Dell PowerMax iSCSI and FC drivers

The Dell PowerMax drivers, PowerMaxISCSIDriver and PowerMaxFCDriver, support the use of Dell PowerMax and VMAX storage arrays with the Cinder Block Storage project. They both provide equivalent functions and differ only in support for their respective host attachment methods.

The drivers perform volume operations by communicating with the back-end PowerMax storage management software. They use the Requests HTTP library to communicate with a Unisphere for PowerMax instance, using a RESTAPI interface in the backend to perform PowerMax and VMAX storage operations.

Note

DEPRECATION NOTICE: The VMAX Hybrid series will not be supported from the Z release of OpenStack. Also, any All Flash array running HyperMaxOS 5977 will no longer be supported from the Z release onwards.

Note

While PowerMax will be used throughout this document, it will be used to collectively categorize the following supported arrays, PowerMax 2000, 8000, VMAX All Flash 250F, 450F, 850F and 950F and VMAX-Hybrid.

System requirements and licensing

The Dell PowerMax Cinder driver supports the VMAX-Hybrid series, VMAX All-Flash series and the PowerMax arrays.

The array operating system software, Solutions Enabler 9.2.2 series, and Unisphere for PowerMax 9.2.2 series are required to run Dell PowerMax Cinder driver for the Wallaby release. Please refer to support-matrix-table for the support matrix of previous OpenStack versions.

Download Solutions Enabler and Unisphere from the Dell’s support web site (login is required). See the Dell Solutions Enabler 9.2.2 Installation and Configuration Guide and Dell Unisphere for PowerMax Installation Guide at the Dell Support site.

Note

At the time each OpenStack release, support-matrix-table was the recommended PowerMax management software and OS combinations. Please reach out your local PowerMax representative to see if these versions are still valid.

PowerMax Management software and OS for OpenStack release

OpenStack

Unisphere for PowerMax

PowerMax OS

Xena

9.2.2

5978.711

Wallaby

9.2.1

5978.711

Victoria

9.2.x

5978.669

Ussuri

9.1.x

5978.479

Train

9.1.x

5978.444

Stein

9.0.x

5978.221

Note

A Hybrid array can only run HyperMax OS 5977, and is still supported until the Z release of OpenStack. Some functionality will not be available in older versions of the OS. If in any doubt, please contact your local PowerMax representative.

Note

Newer versions of Unisphere for PowerMax and PowerMax OS are not retrospectively tested on older versions of OpenStack. If it is necessary to upgrade, the older REST endpoints will be used. For example, in Ussuri, if upgrading to Unisphere for PowerMax 9.2, the older 91 endpoints will be used.

Required PowerMax software suites for OpenStack

The storage system requires a Unisphere for PowerMax (SMC) eLicense.

PowerMax

There are two licenses for the PowerMax 2000 and 8000:

  • Essentials software package

  • Pro software package

The Dell PowerMax cinder driver requires the Pro software package.

All Flash

For full functionality including SRDF for the VMAX All Flash, the FX package, or the F package plus the SRDF a la carte add on is required.

Hybrid

There are five Dell Software Suites sold with the VMAX-Hybrid arrays:

  • Base Suite

  • Advanced Suite

  • Local Replication Suite

  • Remote Replication Suite

  • Total Productivity Pack

The Dell PowerMax Cinder driver requires the Advanced Suite and the Local Replication Suite or the Total Productivity Pack (it includes the Advanced Suite and the Local Replication Suite) for the VMAX Hybrid.

Using PowerMax Remote Replication functionality will also require the Remote Replication Suite.

Note

Each are licensed separately. For further details on how to get the relevant license(s), reference eLicensing Support below.

eLicensing support

To activate your entitlements and obtain your PowerMax license files, visit the Service Center on Dell Support, as directed on your License Authorization Code (LAC) letter emailed to you.

  • For help with missing or incorrect entitlements after activation (that is, expected functionality remains unavailable because it is not licensed), contact your EMC account representative or authorized reseller.

  • For help with any errors applying license files through Solutions Enabler, contact the Dell Customer Support Center.

  • If you are missing a LAC letter or require further instructions on activating your licenses through the Online Support site, contact EMC’s worldwide Licensing team at licensing@emc.com or call:

    North America, Latin America, APJK, Australia, New Zealand: SVC4EMC (800-782-4362) and follow the voice prompts.

    EMEA: +353 (0) 21 4879862 and follow the voice prompts.

PowerMax for OpenStack Cinder customer support

If you require help or assistance with PowerMax and Cinder please open a Service Request (SR) through standard support channels at Dell Support. When opening a SR please include the following information:

  • Array Model & uCode level

  • Unisphere for PowerMax version

  • Solutions Enabler Version

  • OpenStack host Operating System (Ubuntu, RHEL, etc.)

  • OpenStack version (Usurri, Train, etc.)

  • PowerMax for Cinder driver version, this can be located in the comments in the PowerMax driver file: {cinder_install_dir}/cinder/volume/drivers/dell_emc/powermax/fc.py

  • Cinder logs

  • Detailed description of the issue you are encountering

Supported operations

PowerMax drivers support these operations:

  • Create, list, delete, attach, and detach volumes

  • Create, list, and delete volume snapshots

  • Copy an image to a volume

  • Copy a volume to an image

  • Clone a volume

  • Extend a volume

  • Retype a volume (Host and storage assisted volume migration)

  • Create a volume from a snapshot

  • Create and delete generic volume group

  • Create and delete generic volume group snapshot

  • Modify generic volume group (add and remove volumes)

  • Create generic volume group from source

  • Live Migration

  • Volume replication SRDF/S, SRDF/A and SRDF Metro

  • Quality of service (QoS)

  • Manage and unmanage volumes and snapshots

  • List Manageable Volumes/Snapshots

  • Backup create, delete, list, restore and show

PowerMax drivers also support the following features:

  • Dynamic masking view creation

  • Dynamic determination of the target iSCSI IP address

  • iSCSI multipath support

  • Oversubscription

  • Service Level support

  • SnapVX support

  • Compression support(All Flash and PowerMax)

  • Deduplication support(PowerMax)

  • CHAP Authentication

  • Multi-attach support

  • Volume Metadata in logs

  • Encrypted Volume support

  • Extending attached volume

  • Replicated volume retype support

  • Retyping attached(in-use) volume

  • Unisphere High Availability(HA) support

  • Online device expansion of a metro device

  • Rapid TDEV deallocation of deletes

  • Multiple replication devices

  • PowerMax array and storage group tagging

  • Short host name and port group templates

  • Snap id support

  • Seamless Live Migration from SMI-S support

  • Port group & port performance load balancing

Note

In certain cases, when creating a volume from a source snapshot or source volume, subsequent operations using the volumes may fail due to a missing snap_name exception. A manual refresh on the connected Unisphere instance or waiting until another operation automatically refreshes the connected Unisphere instance, will alleviate this issue.

PowerMax naming conventions

Note

shortHostName will be altered using the following formula, if its length exceeds 16 characters. This is because the storage group and masking view names cannot exceed 64 characters:

if len(shortHostName) > 16:
    1. Perform md5 hash on the shortHostName
    2. Convert output of 1. to hex
    3. Take last 6 characters of shortHostName and append output of 2.
    4. If the length of output of 3. exceeds 16 characters, join the
       first 8 characters and last 8 characters.

Note

portgroup_name will be altered using the following formula, if its length exceeds 12 characters. This is because the storage group and masking view names cannot exceed 64 characters:

if len(portgroup_name) > 12:
    1. Perform md5 hash on the portgroup_name
    2. Convert output of 1. to hex
    3. Take last 6 characters of portgroup_name and append output of 2.
    4. If the length of output of 3. exceeds 12 characters, join the
       first 6 characters and last 6 characters.

Masking view names

Masking views are dynamically created by the PowerMax FC and iSCSI drivers using the following naming conventions. [protocol] is either I for volumes attached over iSCSI or F for volumes attached over Fibre Channel.

OS-[shortHostName]-[protocol]-[portgroup_name]-MV

Initiator group names

For each host that is attached to PowerMax volumes using the drivers, an initiator group is created or re-used (per attachment type). All initiators of the appropriate type known for that host are included in the group. At each new attach volume operation, the PowerMax driver retrieves the initiators (either WWNNs or IQNs) from OpenStack and adds or updates the contents of the Initiator Group as required. Names are of the following format. [protocol] is either I for volumes attached over iSCSI or F for volumes attached over Fibre Channel.

OS-[shortHostName]-[protocol]-IG

Note

Hosts attaching to OpenStack managed PowerMax storage cannot also attach to storage on the same PowerMax that are not managed by OpenStack.

FA port groups

PowerMax array FA ports to be used in a new masking view are retrieved from the port group provided as the extra spec on the volume type, or chosen from the list provided in the Dell configuration file.

Storage group names

As volumes are attached to a host, they are either added to an existing storage group (if it exists) or a new storage group is created and the volume is then added. Storage groups contain volumes created from a pool, attached to a single host, over a single connection type (iSCSI or FC). [protocol] is either I for volumes attached over iSCSI or F for volumes attached over Fibre Channel. PowerMax Cinder driver utilizes cascaded storage groups - a parent storage group which is associated with the masking view, which contains child storage groups for each configured SRP/slo/workload/compression-enabled or disabled/replication-enabled or disabled combination.

PowerMax, VMAX All Flash and VMAX-Hybrid

Parent storage group:

OS-[shortHostName]-[protocol]-[portgroup_name]-SG

Child storage groups:

OS-[shortHostName]-[SRP]-[ServiceLevel/Workload]-[portgroup_name]-CD-RE

Note

CD and RE are only set if compression is explicitly disabled or replication explicitly enabled. See the compression 11. All Flash compression support and replication Volume replication support sections below.

Note

For VMAX All Flash with PowerMax OS (5978) or greater, workload if set will be ignored and set to NONE.

Replication storage group naming conventions

Default storage group

Attached child storage group

Management Group

Replication Type

OS-[SRP]-[SL]-[WL]-SG

OS-[HOST]-[SRP]-[SL/WL]-[PG]

N/A

None

OS-[SRP]-[SL]-[WL]-RE-SG

OS-[HOST]-[SRP]-[SL/WL]-[PG]-RE

N/A

Synchronous

OS-[SRP]-[SL]-[WL]-RA-SG

OS-[HOST]-[SRP]-[SL/WL]-[PG]-RA

OS-[RDFG]-Asynchronous-rdf-sg

Asynchronous

OS-[SRP]-[SL]-[WL]-RM-SG

OS-[HOST]-[SRP]-[SL/WL]-[PG]-RM

OS-[RDFG]-Metro-rdf-sg

Metro

PowerMax driver integration

1. Prerequisites

  1. Download Solutions Enabler from Dell Support and install it.

    You can install Solutions Enabler on a non-OpenStack host. Supported platforms include different flavors of Windows, Red Hat, and SUSE Linux. Solutions Enabler can be installed on a physical server, or as a Virtual Appliance (a VMware ESX server VM). Additionally, starting with HYPERMAX OS Q3 2015, you can manage VMAX3 arrays using the Embedded Management (eManagement) container application. See the Dell Solutions Enabler 9.2.1 Installation and Configuration Guide on Dell Support for more details.

    Note

    You must discover storage arrays before you can use the PowerMax drivers. Follow instructions in Dell Solutions Enabler 9.2.1 Installation and Configuration Guide on Dell Support for more details.

  2. Download Unisphere from Dell Support and install it.

    Unisphere can be installed in local, remote, or embedded configurations - i.e., on the same server running Solutions Enabler; on a server connected to the Solutions Enabler server; or using the eManagement container application (containing Solutions Enabler and Unisphere for PowerMax). See Dell Solutions Enabler 9.2.1 Installation and Configuration Guide at Dell Support.

2. FC zoning with PowerMax

Zone Manager is required when there is a fabric between the host and array. This is necessary for larger configurations where pre-zoning would be too complex and open-zoning would raise security concerns.

3. iSCSI with PowerMax

  • Make sure the open-iscsi package (or distro equivalent) is installed on all Compute nodes.

Note

You can only ping the PowerMax iSCSI target ports when there is a valid masking view. An attach operation creates this masking view.

4. Configure block storage in cinder.conf

Description of PowerMax configuration options

Configuration option = Default value

Description

initiator_check = False

(Boolean) Use this value to enable the initiator_check.

interval = 3

(Integer) Use this value to specify length of the interval in seconds.

load_balance = False

(Boolean) Enable/disable load balancing for a PowerMax backend.

load_balance_real_time = False

(Boolean) Enable/disable real-time performance metrics for Port level load balancing for a PowerMax backend.

load_data_format = Avg

(String) Performance data format, not applicable for real-time metrics. Available options are “avg” and “max”.

load_look_back = 60

(Integer) How far in minutes to look back for diagnostic performance metrics in load calculation, minimum of 0 maximum of 1440 (24 hours).

load_look_back_real_time = 1

(Integer) How far in minutes to look back for real-time performance metrics in load calculation, minimum of 1 maximum of 10.

port_group_load_metric = PercentBusy

(String) Metric used for port group load calculation.

port_load_metric = PercentBusy

(String) Metric used for port load calculation.

powermax_array = None

(String) Serial number of the array to connect to.

powermax_array_tag_list = None

(List of String) List of user assigned name for storage array.

powermax_port_group_name_template = portGroupName

(String) User defined override for port group name.

powermax_port_groups = None

(List of String) List of port groups containing frontend ports configured prior for server connection.

powermax_service_level = None

(String) Service level to use for provisioning storage. Setting this as an extra spec in pool_name is preferable.

powermax_short_host_name_template = shortHostName

(String) User defined override for short host name.

powermax_srp = None

(String) Storage resource pool on array to use for provisioning.

retries = 200

(Integer) Use this value to specify number of retries.

u4p_failover_autofailback = True

(Boolean) If the driver should automatically failback to the primary instance of Unisphere when a successful connection is re-established.

u4p_failover_backoff_factor = 1

(Integer) A backoff factor to apply between attempts after the second try (most errors are resolved immediately by a second try without a delay). Retries will sleep for: {backoff factor} * (2 ^ ({number of total retries} - 1)) seconds.

u4p_failover_retries = 3

(Integer) The maximum number of retries each connection should attempt. Note, this applies only to failed DNS lookups, socket connections and connection timeouts, never to requests where data has made it to the server.

u4p_failover_target = None

(Dict of String) Dictionary of Unisphere failover target info.

u4p_failover_timeout = 20.0

(Integer) How long to wait for the server to send data before giving up.

vmax_workload = None

(String) Workload, setting this as an extra spec in pool_name is preferable.

Note

san_api_port is 8443 by default but can be changed if necessary. For the purposes of this documentation the default is assumed so the tag will not appear in any of the cinder.conf extracts below.

Note

PowerMax PortGroups must be pre-configured to expose volumes managed by the array. Port groups can be supplied in cinder.conf, or can be specified as an extra spec storagetype:portgroupname on a volume type. If a port group is set on a volume type as an extra specification it takes precedence over any port groups set in cinder.conf. For more information on port and port group selection please see the section port group & port load balancing.

Note

PowerMax SRP cannot be changed once configured and in-use. SRP renaming on the PowerMax array is not supported.

Note

Service Level can be added to cinder.conf when the backend is the default case and there is no associated volume type. This not a recommended configuration as it is too restrictive. Workload is NONE for PowerMax and any All Flash with PowerMax OS (5978) or greater.

PowerMax parameter

cinder.conf parameter

Default

Required

ServiceLevel

powermax_service_level

None

No

To configure PowerMax block storage, add the following entries to /etc/cinder/cinder.conf:

enabled_backends = CONF_GROUP_ISCSI, CONF_GROUP_FC

[CONF_GROUP_ISCSI]
volume_driver = cinder.volume.drivers.dell_emc.powermax.iscsi.PowerMaxISCSIDriver
volume_backend_name = POWERMAX_ISCSI
powermax_port_groups = [OS-ISCSI-PG]
san_ip = 10.10.10.10
san_login = my_username
san_password = my_password
powermax_array = 000123456789
powermax_srp = SRP_1


[CONF_GROUP_FC]
volume_driver = cinder.volume.drivers.dell_emc.powermax.fc.PowerMaxFCDriver
volume_backend_name = POWERMAX_FC
powermax_port_groups = [OS-FC-PG]
san_ip = 10.10.10.10
san_login = my_username
san_password = my_password
powermax_array = 000123456789
powermax_srp = SRP_1

In this example, two back-end configuration groups are enabled: CONF_GROUP_ISCSI and CONF_GROUP_FC. Each configuration group has a section describing unique parameters for connections, drivers and the volume_backend_name.

5. SSL support

  1. Get the CA certificate of the Unisphere server. This pulls the CA cert file and saves it as .pem file:

    # openssl s_client -showcerts \
                       -connect my_unisphere_host:8443 \
                       </dev/null 2>/dev/null \
                       | openssl x509 -outform PEM > my_unisphere_host.pem
    

    Where my_unisphere_host is the hostname of the unisphere instance and my_unisphere_host.pem is the name of the .pem file.

  2. Add this path to cinder.conf under the PowerMax backend stanza and set SSL verify to True

    driver_ssl_cert_verify = True
    driver_ssl_cert_path = /path/to/my_unisphere_host.pem
    

    OR follow the steps 3-6 below if you would like to add the CA cert to the system certificate bundle instead of specifying the path to cert:

  3. OPTIONAL: Copy the .pem cert to the system certificate directory and convert to .crt:

    # cp my_unisphere_host.pem /usr/share/ca-certificates/ca_cert.crt
    
  4. OPTIONAL: Update CA certificate database with the following command. Ensure you select to enable the cert from step 3 when prompted:

    # sudo dpkg-reconfigure ca-certificates
    
  5. OPTIONAL: Set a system environment variable to tell the Requests library to use the system cert bundle instead of the default Certifi bundle:

    # export REQUESTS_CA_BUNDLE = /etc/ssl/certs/ca-certificates.crt
    
  6. OPTIONAL: Set cert verification to True under the PowerMax backend stanza in cinder.conf:

    # driver_ssl_cert_verify = True
    
  7. Ensure driver_ssl_cert_verify is set to True in cinder.conf backend stanzas if steps 3-6 are followed, otherwise ensure both driver_ssl_cert_path and driver_ssl_cert_verify are set in cinder.conf backend stanzas.

6. Create volume types

Once cinder.conf has been updated, Openstack CLI commands need to be issued in order to create and associate OpenStack volume types with the declared volume_backend_names.

Additionally, each volume type will need an associated pool_name - an extra specification indicating the service level/ workload combination to be used for that volume type.

Note

The pool_name is an additional property which has to be set and is of the format: <ServiceLevel>+<SRP>+<Array ID>. This can be obtained from the output of the cinder get-pools--detail. Workload is NONE for PowerMax or any All Flash with PowerMax OS (5978) or greater.

There is also the option to assign a port group to a volume type by setting the storagetype:portgroupname extra specification.

$ openstack volume type create POWERMAX_ISCSI_SILVER
$ openstack volume type set --property volume_backend_name=ISCSI_backend \
                            --property pool_name=Silver+SRP_1+000123456789 \
                            --property storagetype:portgroupname=OS-PG2 \
                            POWERMAX_ISCSI_SILVER
$ openstack volume type create POWERMAX_FC_DIAMOND
$ openstack volume type set --property volume_backend_name=FC_backend \
                            --property pool_name=Gold+SRP_1+000123456789 \
                            --property storagetype:portgroupname=OS-PG1 \
                            POWERMAX_FC_GOLD

By issuing these commands, the Block Storage volume type POWERMAX_ISCSI_SILVER is associated with the ISCSI_backend, a Silver Service Level.

The type POWERMAX_FC_DIAMOND is associated with the FC_backend, a Diamond Service Level.

The ServiceLevel manages the underlying storage to provide expected performance. Setting the ServiceLevel to None means that non-FAST managed storage groups will be created instead (storage groups not associated with any service level).

openstack volume type set --property pool_name=None+SRP_1+000123456789

Note

PowerMax and VMAX-Hybrid support Diamond, Platinum, Gold, Silver, Bronze, Optimized, and None service levels. VMAX All Flash running HyperMax OS (5977) supports Diamond and None. VMAX-Hybrid and All Flash support DSS_REP, DSS, OLTP_REP, OLTP, and None workloads, the latter up until ucode 5977. Please refer to Stein PowerMax online documentation if you wish to use workload. There is no support for workloads in PowerMax OS (5978) or greater. These will be silently ignored if set for VMAX All-Flash arrays which have been upgraded to PowerMax OS (5988).

7. Interval and retries

By default, interval and retries are 3 seconds and 200 retries respectively. These determine how long (interval) and how many times (retries) a user is willing to wait for a single Rest call, 3*200=600seconds. Depending on usage, these may need to be overridden by the user in cinder.conf. For example, if performance is a factor, then the interval should be decreased to check the job status more frequently, and if multiple concurrent provisioning requests are issued then retries should be increased so calls will not timeout prematurely.

In the example below, the driver checks every 3 seconds for the status of the job. It will continue checking for 200 retries before it times out.

Add the following lines to the PowerMax backend in cinder.conf:

[CONF_GROUP_ISCSI]
volume_driver = cinder.volume.drivers.dell_emc.powermax.iscsi.PowerMaxISCSIDriver
volume_backend_name = POWERMAX_ISCSI
powermax_port_groups = [OS-ISCSI-PG]
san_ip = 10.10.10.10
san_login = my_username
san_password = my_password
powermax_array = 000123456789
powermax_srp = SRP_1
interval = 1
retries = 700

8. CHAP authentication support

This supports one-way initiator CHAP authentication functionality into the PowerMax backend. With CHAP one-way authentication, the storage array challenges the host during the initial link negotiation process and expects to receive a valid credential and CHAP secret in response. When challenged, the host transmits a CHAP credential and CHAP secret to the storage array. The storage array looks for this credential and CHAP secret which stored in the host initiator’s initiator group (IG) information in the ACLX database. Once a positive authentication occurs, the storage array sends an acceptance message to the host. However, if the storage array fails to find any record of the credential/secret pair, it sends a rejection message, and the link is closed.

Assumptions, restrictions and prerequisites

  1. The host initiator IQN is required along with the credentials the host initiator will use to log into the storage array with. The same credentials should be used in a multi node system if connecting to the same array.

  2. Enable one-way CHAP authentication for the iSCSI initiator on the storage array using SYMCLI. Template and example shown below. For the purpose of this setup, the credential/secret used would be my_username/my_password with iSCSI initiator of iqn.1991-05.com.company.lcseb130

    # symaccess -sid <SymmID> -iscsi <iscsi> \
                {enable chap | disable chap | set chap} \
                 -cred <Credential> -secret <Secret>
    
    # symaccess -sid 128 \
                -iscsi iqn.1991-05.com.company.lcseb130 \
                set chap -cred my_username -secret my_password
    

Settings and configuration

  1. Set the configuration in the PowerMax backend group in cinder.conf using the following parameters and restart cinder.

    Configuration options

    Value required for CHAP

    Required for CHAP

    use_chap_auth

    True

    Yes

    chap_username

    my_username

    Yes

    chap_password

    my_password

    Yes

    [POWERMAX_ISCSI]
    volume_driver = cinder.volume.drivers.dell_emc.powermax.iscsi.PowerMaxISCSIDriver
    volume_backend_name = POWERMAX_ISCSI
    san_ip = 10.10.10.10
    san_login = my_u4v_username
    san_password = my_u4v_password
    powermax_srp = SRP_1
    powermax_array = 000123456789
    powermax_port_groups = [OS-ISCSI-PG]
    use_chap_auth = True
    chap_username = my_username
    chap_password = my_password
    

Usage

  1. Using SYMCLI, enable CHAP authentication for a host initiator as described above, but do not set use_chap_auth, chap_username or chap_password in cinder.conf. Create a bootable volume.

    openstack volume create --size 1 \
                            --image <image_name> \
                            --type <volume_type> \
                            test
    
  2. Boot instance named test_server using the volume created above:

    openstack server create --volume test \
                            --flavor m1.small \
                            --nic net-id=private \
                            test_server
    
  3. Verify the volume operation succeeds but the boot instance fails as CHAP authentication fails.

  4. Update cinder.conf with use_chap_auth set to true and chap_username and chap_password set with the correct credentials.

  5. Rerun openstack server create

  6. Verify that the boot instance operation ran correctly and the volume is accessible.

  7. Verify that both the volume and boot instance operations ran successfully and the user is able to access the volume.

9. QoS (Quality of Service) support

Quality of service (QoS) has traditionally been associated with network bandwidth usage. Network administrators set limitations on certain networks in terms of bandwidth usage for clients. This enables them to provide a tiered level of service based on cost. The Nova/Cinder QoS offer similar functionality based on volume type setting limits on host storage bandwidth per service offering. Each volume type is tied to specific QoS attributes some of which are unique to each storage vendor. In the hypervisor, the QoS limits the following:

  • Limit by throughput - Total bytes/sec, read bytes/sec, write bytes/sec

  • Limit by IOPS - Total IOPS/sec, read IOPS/sec, write IOPS/sec

QoS enforcement in Cinder is done either at the hyper-visor (front-end), the storage subsystem (back-end), or both. This section focuses on QoS limits that are enforced by either the PowerMax backend and the hyper-visor front end interchangeably or just back end (Vendor Specific). The PowerMax driver offers support for Total bytes/sec limit in throughput and Total IOPS/sec limit of IOPS.

The PowerMax driver supports the following attributes that are front end/back end agnostic

  • total_iops_sec - Maximum IOPs (in I/Os per second). Valid values range from 100 IO/Sec to 100000 IO/sec.

  • total_bytes_sec - Maximum bandwidth (throughput) in bytes per second. Valid values range from 1048576 bytes (1MB) to 104857600000 bytes (100,000MB)

The PowerMax driver offers the following attribute that is vendor specific to the PowerMax and dependent on the total_iops_sec and/or total_bytes_sec being set.

  • Dynamic Distribution - Enables/Disables dynamic distribution of host I/O limits. Possible values are:

    • Always - Enables full dynamic distribution mode. When enabled, the configured host I/O limits will be dynamically distributed across the configured ports, thereby allowing the limits on each individual port to adjust to fluctuating demand.

    • OnFailure - Enables port failure capability. When enabled, the fraction of configured host I/O limits available to a configured port will adjust based on the number of ports currently online.

    • Never - Disables this feature (Default).

USE CASE 1 - Default values

Prerequisites - PowerMax

  • Host I/O Limit (MB/Sec) - No Limit

  • Host I/O Limit (IO/Sec) - No Limit

  • Set Dynamic Distribution - N/A

Prerequisites - Block Storage (Cinder) back-end (storage group)

Key

Value

total_iops_sec

500

total_bytes_sec

104857600 (100MB)

DistributionType

Always

  1. Create QoS Specs with the prerequisite values above:

    $ openstack volume qos create --consumer back-end \
                                  --property total_iops_sec=500 \
                                  --property total_bytes_sec=104857600 \
                                  --property DistributionType=Always \
                                  my_qos
    
  2. Associate QoS specs with specified volume type:

    $ openstack volume qos associate my_qos my_volume_type
    
  3. Create volume with the volume type indicated above:

    $ openstack volume create --size 1 --type my_volume_type my_volume
    

Outcome - PowerMax (storage group)

  • Host I/O Limit (MB/Sec) - 100

  • Host I/O Limit (IO/Sec) - 500

  • Set Dynamic Distribution - Always

Outcome - Block Storage (Cinder)

Volume is created against volume type and QoS is enforced with the parameters above.

USE CASE 2 - Pre-set limits

Prerequisites - PowerMax

  • Host I/O Limit (MB/Sec) - 2000

  • Host I/O Limit (IO/Sec) - 2000

  • Set Dynamic Distribution - Never

Prerequisites - Block Storage (Cinder) back-end (storage group)

Key

Value

total_iops_sec

500

total_bytes_sec

104857600 (100MB)

DistributionType

Always

  1. Create QoS specifications with the prerequisite values above. The consumer in this use case is both for front-end and back-end:

    $ openstack volume qos create --consumer back-end \
                                  --property total_iops_sec=500 \
                                  --property total_bytes_sec=104857600 \
                                  --property DistributionType=Always \
                                  my_qos
    
  2. Associate QoS specifications with specified volume type:

    $ openstack volume qos associate my_qos my_volume_type
    
  3. Create volume with the volume type indicated above:

    $ openstack volume create --size 1 --type my_volume_type my_volume
    
  4. Attach the volume created in step 3 to an instance

    $ openstack server add volume my_instance my_volume
    

Outcome - PowerMax (storage group)

  • Host I/O Limit (MB/Sec) - 100

  • Host I/O Limit (IO/Sec) - 500

  • Set Dynamic Distribution - Always

Outcome - Block Storage (Cinder)

Volume is created against volume type and QoS is enforced with the parameters above.

Outcome - Hypervisor (Nova)

Libvirt includes an extra xml flag within the <disk> section called iotune that is responsible for rate limitation. To confirm that, first get the OS-EXT-SRV-ATTR:instance_name value of the server instance, for example instance-00000003.

$ openstack server show <serverid>

+-------------------------------------+-----------------------------------------------------------------+
| Field                               | Value                                                           |
+-------------------------------------+-----------------------------------------------------------------+
| OS-DCF:diskConfig                   | AUTO                                                            |
| OS-EXT-AZ:availability_zone         | nova                                                            |
| OS-EXT-SRV-ATTR:host                | myhost                                                          |
| OS-EXT-SRV-ATTR:hypervisor_hostname | myhost                                                          |
| OS-EXT-SRV-ATTR:instance_name       | instance-00000003                                               |
| OS-EXT-STS:power_state              | Running                                                         |
| OS-EXT-STS:task_state               | None                                                            |
| OS-EXT-STS:vm_state                 | active                                                          |
| OS-SRV-USG:launched_at              | 2017-11-02T08:15:42.000000                                      |
| OS-SRV-USG:terminated_at            | None                                                            |
| accessIPv4                          |                                                                 |
| accessIPv6                          |                                                                 |
| addresses                           | private=fd21:99c2:73f3:0:f816:3eff:febe:30ed, 10.0.0.3          |
| config_drive                        |                                                                 |
| created                             | 2017-11-02T08:15:34Z                                            |
| flavor                              | m1.tiny (1)                                                     |
| hostId                              | e7b8312581f9fbb8508587d45c0b6fb4dc86102c632ed1f3a6a49d42        |
| id                                  | 0ef0ff4c-dbda-4dc7-b8ed-45d2fc2f31db                            |
| image                               | cirros-0.3.5-x86_64-disk (b7c220f5-2408-4296-9e58-fc5a41cb7e9d) |
| key_name                            | myhostname                                                      |
| name                                | myhosthame                                                      |
| progress                            | 0                                                               |
| project_id                          | bae4b97a0d8b42c28a5add483981e5db                                |
| properties                          |                                                                 |
| security_groups                     | name='default'                                                  |
| status                              | ACTIVE                                                          |
| updated                             | 2017-11-02T08:15:42Z                                            |
| user_id                             | 7bccf456740546799a7e20457f13c38b                                |
| volumes_attached                    |                                                                 |
+-------------------------------------+-----------------------------------------------------------------+

We then run the following command using the OS-EXT-SRV-ATTR:instance_name retrieved above.

$ virsh dumpxml instance-00000003 | grep -1 "total_bytes_sec\|total_iops_sec"

The output of the command contains the XML below. It is found between the <disk> start and end tag.

<iotune>
   <total_bytes_sec>104857600</total_bytes_sec>
   <total_iops_sec>500</total_iops_sec>
</iotune>

USE CASE 3 - Pre-set limits

Prerequisites - PowerMax

  • Host I/O Limit (MB/Sec) - 100

  • Host I/O Limit (IO/Sec) - 500

  • Set Dynamic Distribution - Always

Prerequisites - Block Storage (Cinder) back end (storage group)

Key

Value

total_iops_sec

500

total_bytes_sec

104857600 (100MB)

DistributionType

OnFailure

  1. Create QoS specifications with the prerequisite values above:

    $ openstack volume qos create --consumer back-end \
                                  --property total_iops_sec=500 \
                                  --property total_bytes_sec=104857600 \
                                  --property DistributionType=OnFailure \
                                  my_qos
    
  2. Associate QoS specifications with specified volume type:

    $ openstack volume qos associate my_qos my_volume_type
    
  3. Create volume with the volume type indicated above:

    $ openstack volume create --size 1 --type my_volume_type my_volume
    

Outcome - PowerMax (storage group)

  • Host I/O Limit (MB/Sec) - 100

  • Host I/O Limit (IO/Sec) - 500

  • Set Dynamic Distribution - OnFailure

Outcome - Block Storage (Cinder)

Volume is created against volume type and QOS is enforced with the parameters above.

USE CASE 4 - Default values

Prerequisites - PowerMax

  • Host I/O Limit (MB/Sec) - No Limit

  • Host I/O Limit (IO/Sec) - No Limit

  • Set Dynamic Distribution - N/A

Prerequisites - Block Storage (Cinder) back end (storage group)

Key

Value

DistributionType

Always

  1. Create QoS specifications with the prerequisite values above:

    $ openstack volume qos create --consumer back-end \
                                  --property DistributionType=Always \
                                  my_qos
    
  2. Associate QoS specifications with specified volume type:

    $ openstack volume qos associate my_qos my_volume_type
    
  3. Create volume with the volume type indicated above:

    $ openstack volume create --size 1 --type my_volume_type my_volume
    

Outcome - PowerMax (storage group)

  • Host I/O Limit (MB/Sec) - No Limit

  • Host I/O Limit (IO/Sec) - No Limit

  • Set Dynamic Distribution - N/A

Outcome - Block Storage (Cinder)

Volume is created against volume type and there is no QoS change.

10. Multi-pathing support

  • Install open-iscsi on all nodes on your system if on an iSCSI setup.

  • Do not install EMC PowerPath as they cannot co-exist with native multi-path software

  • Multi-path tools must be installed on all Nova compute nodes

On Ubuntu:

# apt-get install multipath-tools      #multipath modules
# apt-get install sysfsutils sg3-utils #file system utilities
# apt-get install scsitools            #SCSI tools

On openSUSE and SUSE Linux Enterprise Server:

# zipper install multipath-tools      #multipath modules
# zipper install sysfsutils sg3-utils #file system utilities
# zipper install scsitools            #SCSI tools

On Red Hat Enterprise Linux and CentOS:

# yum install iscsi-initiator-utils   #ensure iSCSI is installed
# yum install device-mapper-multipath #multipath modules
# yum install sysfsutils sg3-utils    #file system utilities

Multipath configuration file

The multi-path configuration file may be edited for better management and performance. Log in as a privileged user and make the following changes to /etc/multipath.conf on the Compute (Nova) node(s).

devices {
# Device attributed for EMC PowerMax
    device {
            vendor "EMC"
            product "SYMMETRIX"
            path_grouping_policy multibus
            getuid_callout "/lib/udev/scsi_id --page=pre-spc3-83 --whitelisted --device=/dev/%n"
            path_selector "round-robin 0"
            path_checker tur
            features "0"
            hardware_handler "0"
            prio const
            rr_weight uniform
            no_path_retry 6
            rr_min_io 1000
            rr_min_io_rq 1
    }
}

You may need to reboot the host after installing the MPIO tools or restart iSCSI and multi-path services.

On Ubuntu iSCSI:

# service open-iscsi restart
# service multipath-tools restart

On Ubuntu FC

# service multipath-tools restart

On openSUSE, SUSE Linux Enterprise Server, Red Hat Enterprise Linux, and CentOS iSCSI:

# systemctl restart open-iscsi
# systemctl restart multipath-tools

On openSUSE, SUSE Linux Enterprise Server, Red Hat Enterprise Linux, and CentOS FC:

# systemctl restart multipath-tools
$ lsblk
NAME                                       MAJ:MIN RM   SIZE RO TYPE  MOUNTPOINT
sda                                          8:0    0     1G  0 disk
..360000970000196701868533030303235 (dm-6) 252:6    0     1G  0 mpath
sdb                                          8:16   0     1G  0 disk
..360000970000196701868533030303235 (dm-6) 252:6    0     1G  0 mpath
vda                                        253:0    0     1T  0 disk

OpenStack configurations

On Compute (Nova) node, add the following flag in the [libvirt] section of nova.conf and nova-cpu.conf:

volume_use_multipath = True

On Cinder controller node, multi-path for image transfer can be enabled in cinder.conf for each backend section or in [backend_defaults] section as a common configuration for all backends.

use_multipath_for_image_xfer = True

Restart nova-compute and cinder-volume services after the change.

Verify you have multiple initiators available on the compute node for I/O

  1. Create a 3GB PowerMax volume.

  2. Create an instance from image out of native LVM storage or from PowerMax storage, for example, from a bootable volume

  3. Attach the 3GB volume to the new instance:

    # multipath -ll
    mpath102 (360000970000196700531533030383039) dm-3 EMC,SYMMETRIX
    size=3G features='1 queue_if_no_path' hwhandler='0' wp=rw
    '-+- policy='round-robin 0' prio=1 status=active
    33:0:0:1 sdb 8:16 active ready running
    '- 34:0:0:1 sdc 8:32 active ready running
    
  4. Use the lsblk command to see the multi-path device:

    # lsblk
    NAME                                       MAJ:MIN RM   SIZE RO TYPE
    sdb                                          8:0    0     3G  0 disk
    ..360000970000196700531533030383039 (dm-6) 252:6    0     3G  0 mpath
    sdc                                          8:16   0     3G  0 disk
    ..360000970000196700531533030383039 (dm-6) 252:6    0     3G  0 mpath
    vda
    

11. All Flash compression support

On an All Flash array, the creation of any storage group has a compressed attribute by default. Setting compression on a storage group does not mean that all the devices will be immediately compressed. It means that for all incoming writes compression will be considered. Setting compression off on a storage group does not mean that all the devices will be uncompressed. It means all the writes to compressed tracks will make these tracks uncompressed.

Note

This feature is only applicable for All Flash arrays, 250F, 450F, 850F and 950F and PowerMax 2000 and 8000. It was first introduced Solutions Enabler 8.3.0.11 or later and is enabled by default when associated with a Service Level. This means volumes added to any newly created storage groups will be compressed.

Use case 1 - Compression disabled create, attach, detach, and delete volume

  1. Create a new volume type called POWERMAX_COMPRESSION_DISABLED.

  2. Set an extra spec volume_backend_name.

  3. Set a new extra spec storagetype:disablecompression = True.

  4. Create a new volume.

  5. Check in Unisphere or SYMCLI to see if the volume exists in storage group OS-<srp>-<servicelevel>-<workload>-CD-SG, and compression is disabled on that storage group.

  6. Attach the volume to an instance. Check in Unisphere or SYMCLI to see if the volume exists in storage group OS-<shorthostname>-<srp>-<servicelevel/workload>-<portgroup>-CD, and compression is disabled on that storage group.

  7. Detach volume from instance. Check in Unisphere or symcli to see if the volume exists in storage group OS-<srp>-<servicelevel>-<workload>-CD-SG, and compression is disabled on that storage group.

  8. Delete the volume. If this was the last volume in the OS-<srp>-<servicelevel>-<workload>-CD-SG storage group, it should also be deleted.

Use case 2 - Retype from compression disabled to compression enabled

  1. Repeat steps 1-4 of Use case 1.

  2. Create a new volume type. For example POWERMAX_COMPRESSION_ENABLED.

  3. Set extra spec volume_backend_name as before.

  4. Set the new extra spec’s compression as storagetype:disablecompression = False or DO NOT set this extra spec.

  5. Retype from volume type POWERMAX_COMPRESSION_DISABLED to POWERMAX_COMPRESSION_ENABLED.

  6. Check in Unisphere or symcli to see if the volume exists in storage group OS-<srp>-<servicelevel>-<workload>-SG, and compression is enabled on that storage group.

Note

If extra spec storagetype:disablecompression is set on a VMAX-Hybrid, it is ignored because compression is not an available feature on a VMAX-Hybrid.

12. Oversubscription support

Please refer to the official OpenStack over-subscription documentation for further information on using over-subscription with PowerMax.

13. Live migration support

Non-live migration (sometimes referred to simply as ‘migration’). The instance is shut down for a period of time to be moved to another hyper-visor. In this case, the instance recognizes that it was rebooted.

Live migration (or ‘true live migration’). Almost no instance downtime. Useful when the instances must be kept running during the migration. The different types of live migration are:

  • Shared storage-based live migration Both hyper-visors have access to shared storage.

  • Block live migration No shared storage is required. Incompatible with read-only devices such as CD-ROMs and Configuration Drive (config_drive).

  • Volume-backed live migration Instances are backed by volumes rather than ephemeral disk. For PowerMax volume-backed live migration, shared storage is required.

The PowerMax driver supports shared volume-backed live migration.

Architecture

In PowerMax, A volume cannot belong to two or more FAST storage groups at the same time. To get around this limitation we leverage both cascaded storage groups and a temporary non-FAST storage group.

A volume can remain ‘live’ if moved between masking views that have the same initiator group and port groups which preserves the host path.

During live migration, the following steps are performed by the PowerMax driver on the volume:

  1. Within the originating masking view, the volume is moved from the FAST storage group to the non-FAST storage group within the parent storage group.

  2. The volume is added to the FAST storage group within the destination parent storage group of the destination masking view. At this point the volume belongs to two storage groups.

  3. One of two things happen:

    • If the connection to the destination instance is successful, the volume is removed from the non-FAST storage group in the originating masking view, deleting the storage group if it contains no other volumes.

    • If the connection to the destination instance fails, the volume is removed from the destination storage group, deleting the storage group, if empty. The volume is reverted back to the original storage group.

Live migration configuration

Please refer to the official OpenStack documentation on configuring migrations and live migration usage for more information.

Note

OpenStack Oslo uses an open standard for messaging middleware known as AMQP. This messaging middleware (the RPC messaging system) enables the OpenStack services that run on multiple servers to talk to each other. By default, the RPC messaging client is set to timeout after 60 seconds, meaning if any operation you perform takes longer than 60 seconds to complete the operation will timeout and fail with the ERROR message Messaging Timeout: Timed out waiting for a reply to message ID [message_id]

If this occurs, increase the rpc_response_timeout flag value in cinder.conf and nova.conf on all Cinder and Nova nodes and restart the services.

What to change this value to will depend entirely on your own environment, you might only need to increase it slightly, or if your environment is under heavy network load it could need a bit more time than normal. Fine tuning is required here, change the value and run intensive operations to determine if your timeout value matches your environment requirements.

At a minimum please set rpc_response_timeout to 240, but this will need to be raised if high concurrency is a factor. This should be sufficient for all Cinder backup commands also.

System configuration

NOVA-INST-DIR/instances/ (for example, /opt/stack/data/nova/instances) has to be mounted by shared storage. Ensure that NOVA-INST-DIR (set with state_path in the nova.conf file) is the same on all hosts.

  1. Configure your DNS or /etc/hosts and ensure it is consistent across all hosts. Make sure that the three hosts can perform name resolution with each other. As a test, use the ping command to ping each host from one another.

    $ ping HostA
    $ ping HostB
    $ ping HostC
    
  2. Export NOVA-INST-DIR/instances from HostA, and ensure it is readable and writable by the Compute user on HostB and HostC. Please refer to the relevant OS documentation for further details, for example Ubuntu NFS Documentation

  3. On all compute nodes, enable the execute/search bit on your shared directory to allow qemu to be able to use the images within the directories. On all hosts, run the following command:

    $ chmod o+x NOVA-INST-DIR/instances
    

Note

If migrating from compute to controller, make sure to run step two above on the controller node to export the instance directory.

Use case

For our use case shown below, we have three hosts with host names HostA, HostB and HostC. HostA is the controller node while HostB and HostC are the compute nodes. The following were also used in live migration.

  • 2GB bootable volume using the CirrOS image.

  • Instance created using the 2GB volume above with a flavor m1.small using 2048 RAM, 20GB of Disk and 1 VCPU.

  1. Create a bootable volume.

    $ openstack volume create --size 2 \
                              --image cirros-0.3.5-x86_64-disk \
                              --volume_lm_1
    
  2. Launch an instance using the volume created above on HostB.

    $ openstack server create --volume volume_lm_1 \
                              --flavor m1.small \
                              --nic net-id=private \
                              --security-group default \
                              --availability-zone nova:HostB \
                              server_lm_1
    
  3. Confirm on HostB has the instance created by running:

    $ openstack server show server_lm_1 | grep "hypervisor_hostname\|instance_name"
      | OS-EXT-SRV-ATTR:hypervisor_hostname | HostB
      | OS-EXT-SRV-ATTR:instance_name | instance-00000006
    
  4. Confirm, through virsh using the instance_name returned in step 3 (instance-00000006), on HostB that the instance is created using:

    $ virsh list --all
    
    Id   Name                  State
    --------------------------------
    1    instance-00000006     Running
    
  5. Migrate the instance from HostB to HostA with:

    $ openstack server migrate --os-compute-api-version 2.30 \
                               --live-migration --host HostA \
                               server_lm_1
    
  6. Run the command on step 3 above when the instance is back in available status. The hypervisor should be on Host A.

  7. Run the command on Step 4 on Host A to confirm that the instance is created through virsh.

14. Multi-attach support

PowerMax cinder driver supports the ability to attach a volume to multiple hosts/servers simultaneously. Please see the official OpenStack multi-attach documentation for configuration information.

Multi-attach architecture

In PowerMax, a volume cannot belong to two or more FAST storage groups at the same time. This can cause issues when we are attaching a volume to multiple instances on different hosts. To get around this limitation, we leverage both cascaded storage groups and non-FAST storage groups (i.e. a storage group with no service level, workload, or SRP specified).

Note

If no service level is assigned to the volume type, no extra work on the backend is required – the volume is attached to and detached from each host as normal.

Example use case

Volume Multi-attach-Vol-1 (with a multi-attach capable volume type, and associated with a Diamond Service Level) is attached to Instance Multi-attach-Instance-A on HostA. We then issue the command to attach Multi-attach-Vol-1 to Multi-attach-Instance-B on HostB:

  1. In the HostA masking view, the volume is moved from the FAST managed storage group to the non-FAST managed storage group within the parent storage group.

  2. The volume is attached as normal on HostB – i.e., it is added to a FAST managed storage group within the parent storage group of the HostB masking view. The volume now belongs to two masking views, and is exposed to both HostA and HostB.

We then decide to detach the volume from Multi-attach-Instance-B on HostB:

  1. The volume is detached as normal from HostB – i.e., it is removed from the FAST managed storage group within the parent storage group of the HostB masking view – this includes cleanup of the associated elements if required. The volume now belongs to one masking view, and is no longer exposed to HostB.

  2. In the HostA masking view, the volume is returned to the FAST managed storage group from the non-FAST managed storage group within the parent storage group. The non-FAST managed storage group is cleaned up, if required.

15. Volume encryption support

Encryption is supported through the use of OpenStack Barbican. Only front-end encryption is supported, back-end encryption is handled at the hardware level with Data at Rest Encryption (D@RE).

For further information on OpenStack Barbican including setup and configuration please refer to the following official Barbican documentation.

16. Volume metadata

Volume metadata is returned to the user in both the Cinder Volume logs and with volumes and snapshots created in Cinder via the UI or CLI.

16.1 Volume metadata in logs

If debug is enabled in the default section of cinder.conf, PowerMax Cinder driver will log additional volume information in the Cinder volume log, on each successful operation. The facilitates bridging the gap between OpenStack and the Array by tracing and describing the volume from a VMAX/ PowerMax view point.

+------------------------------------+---------------------------------------------------------+
| Key                                | Value                                                   |
+------------------------------------+---------------------------------------------------------+
| service_level                      | Gold                                                    |
| is_compression_disabled            | no                                                      |
| powermax_cinder_driver_version     | 3.2.0                                                   |
| identifier_name                    | OS-819470ab-a6d4-49cc-b4db-6f85e82822b7                 |
| openstack_release                  | 13.0.0.0b3.dev3                                         |
| volume_id                          | 819470ab-a6d4-49cc-b4db-6f85e82822b7                    |
| storage_model                      | PowerMax_8000                                           |
| successful_operation               | delete                                                  |
| default_sg_name                    | OS-DEFAULT_SRP-Gold-NONE-SG                             |
| device_id                          | 01C03                                                   |
| unisphere_for_powermax_version     | V9.0.0.9                                                |
| workload                           | NONE                                                    |
| openstack_version                  | 13.0.0                                                  |
| volume_updated_time                | 2018-08-03 03:13:53                                     |
| platform                           | Linux-4.4.0-127-generic-x86_64-with-Ubuntu-16.04-xenial |
| python_version                     | 2.7.12                                                  |
| volume_size                        | 20                                                      |
| srp                                | DEFAULT_SRP                                             |
| openstack_name                     | 90_Test_Vol56                                           |
| storage_firmware_version           | 5978.143.144                                            |
| serial_number                      | 000123456789                                            |
+------------------------------------+---------------------------------------------------------+

16.2 Metadata in the UI and CLI

By default metadata will be set on all volume and snapshot objects created in Cinder. This information represents the state of the object on the backend PowerMax and will change when volume attributes are changed by performing actions on them such as re-type or attaching to an instance.

demo@openstack-controller:~$ cinder show powermax-volume

+--------------------------------+------------------------------------------------------------+
| Property                       | Value                                                      |
+--------------------------------+------------------------------------------------------------+
| metadata                       | ArrayID : 000123456789                                     |
|                                | ArrayModel : PowerMax_8000                                 |
|                                | CompressionDisabled : False                                |
|                                | Configuration : TDEV                                       |
|                                | DeviceID : 0012F                                           |
|                                | DeviceLabel : OS-d87edb98-60fd-49dd-bb0f-cc388cf6f3f4      |
|                                | Emulation : FBA                                            |
|                                | ReplicationEnabled : False                                 |
|                                | ServiceLevel : Diamond                                     |
|                                | Workload : None                                            |
| name                           | powermax-volume                                            |
+--------------------------------+------------------------------------------------------------+

17. Unisphere High Availability (HA) support

This feature facilitates high availability of Unisphere for PowerMax servers, allowing for one or more backup unisphere instances in the event of a loss in connection to the primary Unisphere instance. The PowerMax driver will cycle through the list of failover instances, trying each until a successful connection is made. The ordering is first in, first out (FIFO), so the first u4p_failover_target specified in cinder.conf will be the first selected, the second u4p_failover_target in cinder.conf will be the second selected, and so on until all failover targets are exhausted.

Requirements

  • All required instances of Unisphere for PowerMax are set up and configured for the array(s)

  • Array(s) are locally registered with the instance of Unisphere that will be used as a failover instance. There are two failover types, local and remote:

    • Local failover - Primary Unisphere is unreachable, failover to secondary local instance of Unisphere to resume normal operations at primary site.

    • Remote failover - Complete loss of primary site so primary instance of Unisphere is unreachable, failover to secondary instance of Unisphere at remote site to resume operations with the R2 array.

Note

Replication must be configured in advance for remote failover to work successfully. Human intervention will also be required to failover from R1 array to R2 array in Cinder using cinder failover-host command (see Volume replication support for replication setup details).

Note

The remote target array must be registered as local to the remote instance of Unisphere

Configuration

The following configuration changes need to be made in cinder.conf under the PowerMax backend stanza in order to support the failover to secondary Unisphere. Cinder services will need to be restarted for changes to take effect.

[POWERMAX_1]
...

u4p_failover_timeout = 30
u4p_failover_retries = 3
u4p_failover_backoff_factor = 1
u4p_failover_autofailback = True
u4p_failover_target = san_ip:10.10.10.12,
                      san_api_port: 8443,
                      san_login:my_username,
                      san_password:my_password,
                      driver_ssl_cert_verify: False,
u4p_failover_target = san_ip:10.10.10.13,
                      san_api_port: 8443
                      san_login:my_username,
                      san_password:my_password,
                      driver_ssl_cert_verify: True,
                      driver_ssl_cert_path: /path/to/my_unisphere_host.pem

Note

u4p_failover_target key value pairs will need to be on the same line (separated by commas) in cinder.conf. They are displayed on separated lines above for readability.

Note

To add more than one Unisphere failover target create additional u4p_failover_target details for the Unisphere instance. These will be cycled through in a first-in, first-out (FIFO) basis, the first failover target in cinder.conf will be the first backup instance of Unisphere used by the PowerMax driver.

18. Rapid TDEV deallocation

The PowerMax driver can now leverage the enhanced volume delete feature-set made available in the PowerMax 5978 Foxtail uCode release. These enhancements allow volume deallocation & deletion to be combined into a single call. Previously, volume deallocation & deletion were split into separate tasks; now a single REST call is dispatched and a response code on the projected outcome of their request is issued rapidly allowing other task execution to proceed without the delay. No additional configuration is necessary, the system will automatically determine when to use either the rapid or legacy compliant volume deletion sequence based on the connected PowerMax array’s metadata.

19. PowerMax online (in-use) device expansion

uCode Level

Supported In-Use Volume Extend Operations

R1 uCode Level

R2 uCode Level

Sync

Async

Metro

5978.711

5978.711

Y

Y

Y

5978.711

5978.669

Y

Y

Y

5978.711

5978.444

Y

Y

Y

5978.711

5978.221

Y

Y

N

5978.669

5978.669

Y

Y

Y

5978.669

5978.444

Y

Y

Y

5978.669

5978.221

Y

Y

N

5978.444

5978.444

Y

Y

Y

5978.444

5978.221

Y

Y

N

5978.221

5978.221

Y

Y

N

Assumptions, restrictions and prerequisites

  • ODE in the context of this document refers to extending a volume where it is in-use, that is, attached to an instance.

  • The allow_extend is only applicable on VMAX-Hybrid arrays or All Flash arrays with HyperMax OS. If included elsewhere, it is ignored.

  • Where one array is a lower uCode than the other, the environment is limited to functionality of that of the lowest uCode level, i.e. if R1 is 5978.444 and R2 is 5978.221, expanding a metro volume is not supported, both R1 and R2 need to be on 5978.444 uCode at a minimum.

20. PowerMax array and storage group tagging

Unisphere for PowerMax 9.1 and later supports tagging of storage groups and arrays, so the user can give their own ‘tag’ for ease of searching and/or grouping.

Assumptions, restrictions and prerequisites

  • The storage group tag(s) is associated with a volume type extra spec key storagetype:storagegrouptags.

  • The array tag is associated with the backend stanza using key powermax_array_tag_list. It expects a list of one or more comma separated values, for example powermax_array_tag_list=[value1,value2, value3]

  • They can be one or more values in a comma separated list.

  • There is a 64 characters limit of letters, numbers, - and _.

  • 8 tags are allowed per storage group and array.

  • Tags cannot be modified once a volume has been created with that volume type. This is an OpenStack constraint.

  • Tags can be modified on the backend stanza, but none will ever be removed, only added.

  • There is no restriction on creating or deleting tags of OpenStack storage groups or arrays outside of OpenStack, for example Unisphere for PowerMax UI. The max number of 8 tags will apply however, as this is a Unisphere for PowerMax limit.

Set a storage group tag on a volume type:

$ openstack volume type set --property storagetype:storagegrouptags=myStorageGroupTag1,myStorageGroupTag2

Set an array tag on the PowerMax backend:

[POWERMAX_ISCSI]
volume_driver = cinder.volume.drivers.dell_emc.powermax.iscsi.PowerMaxISCSIDriver
volume_backend_name = POWERMAX_ISCSI
san_ip = 10.10.10.10
san_login = my_u4v_username
san_password = my_u4v_password
powermax_srp = SRP_1
powermax_array = 000123456789
powermax_port_groups = [OS-ISCSI-PG]
powermax_array_tag_list = [openstack1, openstack2]

21. PowerMax short host name and port group name override

This functionality allows the user to customize the short host name and port group name that are contained in the PowerMax driver storage groups and masking views names. For current functionality please refer to PowerMax naming conventions for more details.

As the storage group name and masking view name are limited to 64 characters the short host name needs to be truncated to 16 characters or less and port group needs to be truncated to 12 characters or less. This functionality offers a little bit more flexibility to determine how these truncated components should look.

Note

Once the port group and short host name have been overridden with any new format, it is not possible to return to the default format or change to another format if any volumes are in an attached state. This is because there is no way to determine the overridden format once powermax_short_host_name_template` or ``powermax_port_group_name_template have been removed or changed.

Assumptions, restrictions, and prerequisites

  • Backward compatibility with old format is preserved.

  • cinder.conf will have 2 new configuration options, short_host_name_template and port_group_name_template.

  • If a storage group, masking view or initiator group in the old naming convention already exists, this remains and any new attaches will use the new naming convention where the label for the short host name and/or port group has been customized by the user.

  • Only the short host name and port group name components can be renamed within the storage group, initiator group and masking view names.

  • If the powermax_short_host_name_template and powermax_port_group_name_template do not adhere to the rules, then the operation will fail early and gracefully with a clear description as to the problem.

  • The templates cannot be changed once volumes have been attached using the new configuration.

  • If only one of the templates are configured, then the other will revert to the default option.

  • The UUID is generated from the MD5 hash of the full short host name and port group name

  • If userdef is used, the onus is on the user to make sure it will be unique among all short host names (controller and compute nodes) and unique among port groups.

Short host name templates

powermax_short_host_name_template

Description

Rule

shortHostName

This is the default option

Existing functionality, if over 16 characters then see PowerMax naming conventions, otherwise short host name

shortHostName[:x])uuid[:x] e.g. shortHostName[:6]uuid[:9]

First x characters of the short host name and x uuid characters created from md5 hash of short host name

Must be less than 16 characters

shortHostName[:x]userdef e.g. shortHostName[:6]-testHost

First x characters of the short host name and a user defined x char name. NB - the responsibility is on the user for uniqueness

Must be less than 16 characters

shortHostName[-x:]uuid[:x] e.g. shortHostName[-6:]uuid[:9]

Last x characters of the short host name and x uuid characters created from md5 hash of short host name

Must be less than 16 characters

shortHostName[-x:]userdef e.g. shortHostName[-6:]-testHost

Last x characters of the short host name and a user defined x char name. NB - the responsibility is on the user for uniqueness

Must be less than 16 characters

Port group name templates

powermax_port_group_name_template

Description

Rule

portGroupName

This is the default option

Existing functionality, if over 12 characters then see PowerMax naming conventions, otherwise port group name

portGroupName[:x])uuid[:x] e.g. portGroupName[:6]uuid[:5]

First x characters of the port group name and x uuid characters created from md5 hash of port group name

Must be less than 12 characters

portGroupName[:x]userdef e.g. portGroupName[:6]-test

First x characters of the port group name and a user defined x char name. NB - the responsibility is on the user for uniqueness

Must be less than 12 characters

portGroupName[-x:]uuid[:x] e.g. portGroupName[-6:]uuid[:5]

Last x characters of the port group name and x uuid characters created from md5 hash of port group name

Must be less than 12 characters

portGroupName[-x:]userdef e.g. portGroupName[-6:]-test

Last x characters of the port group name and a user defined x char name. NB - the responsibility is on the user for uniqueness

Must be less than 12 characters

21. Snap ids replacing generations

Snap ids were introduced to the PowerMax in microcde 5978.669.669 and Unisphere for PowerMax 9.2. Generations existed previously and could cause stale data if deleted out of sequence, even though we locked against this occurence. This happened when the newer generation(s) inherited its deleted predecessors generation number. So in a series of 0, 1, 2 and 3 generations, if generation 1 gets deleted, generation 2 now becomes generation 1 and generation 3 becomes generation 2 and so on down the line. Snap ids are unique to each snapVX and will not change once assigned at creation so out of sequence deletions are no longer an issue. Generations will remain for arrays with microcode less than 5978.669.669.

Cinder supported operations

Volume replication support

Note

A mix of RDF1+TDEV and TDEV volumes should not exist in the same storage group. This can happen on a cleanup operation after breaking the pair and a ‘TDEV’ remains in the storage group on either the local or remote array. If this happens, remove the volume from the storage group so that further replicated volume operations can continue. For example, Remove TDEV from OS-[SRP]-[SL]-[WL]-RA-SG.

Note

Replication storage groups should exist on both local and remote array but never on just one. For example, if OS-[SRP]-[SL]-[WL]-RA-SG exists on local array A it must also exist on remote array B. If this condition does not hold, further replication operations will fail. This applies to management storage groups in the case of Asynchronous and Metro modes also. See Replication storage group naming conventions.

Note

The number of devices in replication storage groups in both local and remote arrays should be same. This also applies to management storage groups in Asynchronous and Metro modes. See Replication storage group naming conventions.

Configure a single replication target

  1. Configure an SRDF group between the chosen source and target arrays for the PowerMax Cinder driver to use. The source array must correspond with the powermax_array entry in cinder.conf.

  2. Select both the director and the ports for the SRDF emulation to use on both sides. Bear in mind that network topology is important when choosing director endpoints. Supported modes are Synchronous, Asynchronous, and Metro.

    Note

    If the source and target arrays are not managed by the same Unisphere server (that is, the target array is remotely connected to server - for example, if you are using embedded management), in the event of a full disaster scenario (i.e. the primary array is completely lost and all connectivity to it is gone), the Unisphere server would no longer be able to contact the target array. In this scenario, the volumes would be automatically failed over to the target array, but administrator intervention would be required to either; configure the target (remote) array as local to the current Unisphere server (if it is a stand-alone server), or enter the details of a second Unisphere server to the cinder.conf, which is locally connected to the target array (for example, the embedded management Unisphere server of the target array), and restart the Cinder volume service.

    Note

    If you are setting up an SRDF/Metro configuration, it is recommended that you configure a Witness or vWitness for bias management. Please see the SRDF Metro Overview & Best Practices guide for more information.

    Note

    The PowerMax Cinder drivers do not support Cascaded SRDF.

    Note

    The transmit idle functionality must be disabled on the R2 array for Asynchronous rdf groups. If this is not disabled it will prevent failover promotion in the event of access to the R1 array being lost.

    # symrdf -sid <sid> -rdfg <rdfg> set rdfa -transmit_idle off
    

    Note

    When creating RDF enabled volumes, if there are existing volumes in the target storage group, all rdf pairs related to that storage group must have the same rdf state i.e. rdf pair states must be consistent across all volumes in a storage group when attempting to create a new replication enabled volume. If mixed rdf pair states are found during a volume creation attempt, an error will be raised by the rdf state validation checks. In this event, please wait until all volumes in the storage group have reached a consistent state.

  3. Enable replication in /etc/cinder/cinder.conf. To enable the replication functionality in PowerMax Cinder driver, it is necessary to create a replication volume-type. The corresponding back-end stanza in cinder.conf for this volume-type must then include a replication_device parameter. This parameter defines a single replication target array and takes the form of a list of key value pairs.

    enabled_backends = POWERMAX_FC_REPLICATION
    [POWERMAX_FC_REPLICATION]
    volume_driver = cinder.volume.drivers.dell_emc.powermax.fc.PowerMaxFCDriver
    san_ip = 10.10.10.10
    san_login = my_u4v_username
    san_password = my_u4v_password
    powermax_srp = SRP_1
    powermax_array = 000123456789
    powermax_port_groups = [OS-FC-PG]
    volume_backend_name = POWERMAX_FC_REPLICATION
    replication_device = target_device_id:000197811111,
                         remote_port_group:os-failover-pg,
                         remote_pool:SRP_1,
                         rdf_group_label: 28_11_07,
                         mode:Metro,
                         metro_use_bias:False,
                         sync_interval:3,
                         sync_retries:200
    

    Note

    replication_device key value pairs will need to be on the same line (separated by commas) in cinder.conf. They are displayed here on separate lines above for improved readability.

    • target_device_id The unique PowerMax array serial number of the target array. For full failover functionality, the source and target PowerMax arrays must be discovered and managed by the same U4V server.

    • remote_port_group The name of a PowerMax port group that has been pre-configured to expose volumes managed by this backend in the event of a failover. Make sure that this port group contains either all FC or all iSCSI port groups (for a given back end), as appropriate for the configured driver (iSCSI or FC).

    • remote_pool The unique pool name for the given target array.

    • rdf_group_label The name of a PowerMax SRDF group that has been pre-configured between the source and target arrays.

    • mode The SRDF replication mode. Options are Synchronous, Asynchronous, and Metro. This defaults to Synchronous if not set.

    • metro_use_bias Flag to indicate if ‘bias’ protection should be used instead of Witness. This defaults to False.

    • sync_interval How long in seconds to wait between intervals for SRDF sync checks during Cinder PowerMax SRDF operations. Default is 3 seconds.

    • sync_retries How many times to retry RDF sync checks during Cinder PowerMax SRDF operations. Default is 200 retries.

    • allow_extend Only applicable to VMAX-Hybrid arrays or All Flash arrays running HyperMax OS (5977). It is a flag for allowing the extension of replicated volumes. To extend a volume in an SRDF relationship, this relationship must first be broken, the R1 device extended, and a new device pair established. If not explicitly set, this flag defaults to False.

      Note

      As the SRDF link must be severed, due caution should be exercised when performing this operation. If absolutely necessary, only one source and target pair should be extended at a time (only only applicable to VMAX-Hybrid arrays or All Flash arrays with HyperMax OS).

  4. Create a replication-enabled volume type. Once the replication_device parameter has been entered in the PowerMax backend entry in the cinder.conf, a corresponding volume type needs to be created replication_enabled property set. See above Create volume types for details.

    # openstack volume type set --property replication_enabled="<is> True" \
                          POWERMAX_FC_REPLICATION
    

    Note

    Service Level and Workload: An attempt will be made to create a storage group on the target array with the same service level and workload combination as the primary. However, if this combination is unavailable on the target (for example, in a situation where the source array is a VMAX-Hybrid, the target array is an All Flash, and an All Flash incompatible service level like Bronze is configured), no service level will be applied.

Configure multiple replication targets

Setting multiple replication devices in cinder.conf allows the use of all the supported replication modes simultaneously. Up to three replication devices can be set, one for each of the replication modes available. An additional volume type extra spec (storagetype:replication_device_backend_id) is then used to determine which replication device should be utilized when attempting to perform an operation on a volume which is replication enabled. All details, guidelines and recommendations set out in the Configure a single replication target section also apply in a multiple replication device scenario.

Multiple replication targets limitations and restrictions:
  1. There can only be one of each replication mode present across all of the replication devices set in cinder.conf.

  2. Details for target_device_id, remote_port_group and remote_pool should be identical across replication devices.

  3. The backend_id and rdf_group_label values must be unique across all replication devices.

Adding additional replication_device to cinder.conf:
  1. Open cinder.conf for editing

  2. If a replication device is already present, add the backend_id key with a value of backend_id_legacy_rep. If this key is already defined, it’s value must be updated to backend_id_legacy_rep.

  3. Add the additional replication devices to the backend stanza. Any additional replication devices must have a backend_id key set. The value of these must not be backend_id_legacy_rep.

Example existing backend stanza pre-multiple replication:

enabled_backends = POWERMAX_FC_REPLICATION

[POWERMAX_FC_REPLICATION]
volume_driver = cinder.volume.drivers.dell_emc.powermax.fc.PowerMaxFCDriver
san_ip = 10.10.10.10
san_login = my_u4v_username
san_password = my_u4v_password
powermax_srp = SRP_1
powermax_array = 000123456789
powermax_port_groups = [OS-FC-PG]
volume_backend_name = POWERMAX_FC_REPLICATION
replication_device = backend_id:id,
                     target_device_id:000197811111,
                     remote_port_group:os-failover-pg,
                     remote_pool:SRP_1,
                     rdf_group_label: 28_11_07,
                     mode:Metro,
                     metro_use_bias:False,
                     sync_interval:3,
                     sync_retries:200

Example updated backend stanza:

enabled_backends = POWERMAX_FC_REPLICATION

[POWERMAX_FC_REPLICATION]
volume_driver = cinder.volume.drivers.dell_emc.powermax.fc.PowerMaxFCDriver
san_ip = 10.10.10.10
san_login = my_u4v_username
san_password = my_u4v_password
powermax_srp = SRP_1
powermax_array = 000123456789
powermax_port_groups = [OS-FC-PG]
volume_backend_name = POWERMAX_FC_REPLICATION
replication_device = backend_id:backend_id_legacy_rep
                     target_device_id:000197811111,
                     remote_port_group:os-failover-pg,
                     remote_pool:SRP_1,
                     rdf_group_label: 28_11_07,
                     mode:Metro,
                     metro_use_bias:False,
                     sync_interval:3,
                     sync_retries:200
replication_device = backend_id:sync-rep-id
                     target_device_id:000197811111,
                     remote_port_group:os-failover-pg,
                     remote_pool:SRP_1,
                     rdf_group_label: 29_12_08,
                     mode:Synchronous,
                     sync_interval:3,
                     sync_retries:200
replication_device = backend_id:async-rep-id
                     target_device_id:000197811111,
                     remote_port_group:os-failover-pg,
                     remote_pool:SRP_1,
                     rdf_group_label: 30_13_09,
                     mode:Asynchronous,
                     sync_interval:3,
                     sync_retries:200

Note

For environments without existing replication devices. The backend_id values can be set to any value for all replication devices. The backend_id_legacy_rep value is only needed when updating a legacy system with an existing replication device to use multiple replication devices.

The additional replication devices defined in cinder.conf will be detected after restarting the cinder volume service.

To specify which replication_device a volume type should use an additional property named storagetype:replication_device_backend_id must be added to the extra specs of the volume type. The id value assigned to the storagetype:replication_device_backend_id key in the volume type must match the backend_id assigned to the replication_device in cinder.conf.

# openstack volume type set \
--property storagetype:replication_device_backend_id="<id>" \
<VOLUME_TYPE>

Note

Specifying which replication device to use is done in addition to the basic replication setup for a volume type seen in Configure a single replication target

Note

In a legacy system where volume types are present that were replication enabled before adding multiple replication devices, the storagetype:replication_device_backend_id should be omitted from any volume type that does/will use the legacy replication_device i.e. when storagetype:replication_device_backend_id is omitted the replication_device with a backend_id of backend_id_legacy_rep will be used.

Volume replication interoperability with other features

Most features are supported, except for the following:

  • Replication Group operations are available for volumes in Synchronous mode only.

  • The Ussuri release of OpenStack supports retyping in-use volumes to and from replication enabled volume types with limited exception of volumes with Metro replication enabled. To retype to a volume-type that is Metro enabled the volume must first be detached then retyped. The reason for this is so the paths from the Nova instance to the Metro R1 & R2 volumes must be initialised, this is not possible on the R2 device whilst a volume is attached.

  • The image volume cache functionality is supported (enabled by setting image_volume_cache_enabled = True), but one of two actions must be taken when creating the cached volume:

    • The first boot volume created on a backend (which will trigger the cached volume to be created) should be the smallest necessary size. For example, if the minimum size disk to hold an image is 5GB, create the first boot volume as 5GB. All subsequent boot volumes are extended to the user specific size.

    • Alternatively, ensure that the allow_extend option in the replication_device parameter is set to True. This is only applicable to VMAX-Hybrid arrays or All Flash array with HyperMax OS.

Failover host

Note

Failover and failback operations are not applicable in Metro configurations.

In the event of a disaster, or where there is required downtime, upgrade of the primary array for example, the administrator can issue the failover host command to failover to the configured target:

# cinder failover-host cinder_host@POWERMAX_FC_REPLICATION

After issuing cinder failover-host Cinder will set the R2 array as the target array for Cinder, however, to get existing instances to use this new array and paths to volumes it is necessary to first shelve Nova instances and then unshelve them, this will effectively restart the Nova instance and re-establish data paths between Nova instances and the volumes on the R2 array.

# nova shelve <server>
# nova unshelve [--availability-zone <availability_zone>] <server>

When a host is in failover mode performing normal volume or snapshot provisioning will not be possible, failover host mode simply provides access to replicated volumes to minimise environment down-time. The primary objective whilst in failover mode should be to get the R1 array back online. When the primary array becomes available again, you can initiate a fail-back using the same failover command and specifying --backend_id default:

# cinder failover-host cinder_host@POWERMAX_FC_REPLICATION --backend_id default

After issuing the failover command to revert to the default backend host it is necessary to re-issue the Nova shelve and unshelve commands to restore the data paths between Nova instances and their corresponding back end volumes. Once reverted to the default backend volume and snapshot provisioning operations can continue as normal.

Failover promotion

Failover promotion can be used to transfer all existing RDF enabled volumes to the R2 array and overwrite any references to the original R1 array. This can be used in the event of total R1 array failure or in other cases where an array transfer is warranted. If the R1 array is online and working and the RDF links are still enabled the failover promotion will automatically delete rdf pairs as necessary. If the R1 array or the link to the R1 array is down, a half deletepair must be issued manually for those volumes during the failover promotion.

  1. Issue failover command:

# cinder failover-host <host>
  1. Enable array promotion:

# cinder failover-host --backend_id=pmax_failover_start_array_promotion <host>
  1. View and re-enable the cinder service

# cinder service-list
# cinder service-enable <host> <binary>
  1. Remove all volumes from volume groups

# cinder --os-volume-api-version 3.13 group-update --remove-volumes <Vol1ID, etc..> <volume_group_name>
  1. Detach all volumes that are attached to instances

# openstack server remove volume <instance_id> <volume_id>

Note

Deleting the instance will call a detach volume for each attached volume. A terminate connection can be issued manually using the following command for volumes that are stuck in the attached state without an instance.

# cinder --os-volume-api-version 3.50 attachment-delete <attachment_id>
  1. Delete all remaining instances

# nova delete <instance_id>
  1. Create new volume types

New volume types must be created with references to the remote array. All new volume types must adhere to the following guidelines:

1. Uses the same workload, SLO & compression setting as the previous R1 volume type.
2. Uses the remote array instead of the primary for its pool name.
3. Uses the same volume_backend_name as the previous volume type.
4. Must not have replication enabled.

Example existing volume type extra specs.

pool_name='Gold+None+SRP_1+000297900330', replication_enabled='<is> True',
storagetype:replication_device_backend_id='async-rep-1', volume_backend_name='POWERMAX_ISCSI_NONE'

Example new volume type extra specs.

pool_name='Gold+None+SRP_1+000197900049', volume_backend_name='POWERMAX_ISCSI_NONE'
  1. Retype volumes to new volume types

Additional checks will be performed during failover promotion retype to ensure workload, compression and slo settings meet the criteria specified above when creating the new volume types.

# cinder retype --migration-policy on-demand <volume> <volume_type>

Note

If the volumes RDF links are offline during this retype then a half deletepair must be performed manually after retype. Please reference section 8.a. below for guidance on this process.

8.a. Retype and RDF half deletepair

In instances where the rdf links are offline and rdf pairs have been set to partitioned state there are additional requirements. In that scenario the following order should be adhered to:

1. Retype all Synchronous volumes.
2. Half_deletepair all Synchronous volumes using the default storage group.
3. Retype all Asynchronous volumes.
4. Half_deletepair all Asynchronous volumes using their management storage group.
5. Retype all Metro volumes.
6. Half_deletepair all Metro volumes using their management storage group.
7. Delete the Asynchronous and Metro management storage groups.

Note

A half deletepair cannot be performed on Metro enabled volumes unless the symforce option has been enabled in the symapi options. In symapi/config/options uncomment and set ‘SYMAPI_ALLOW_RDF_SYMFORCE = True’.

# symrdf -sid <sid> -sg <sg> -rdfg <rdfg> -force -symforce half_deletepair
  1. Issue failback

Issuing the failback command will disable both the failover and promotion flags. Please ensure all volumes have been retyped and all replication pairs have been deleted before issuing this command.

# cinder failover-host --backend_id default <host>
  1. Update cinder.conf

Update the cinder.conf file to include details for the new primary array. For more information please see the Configure block storage in cinder.conf section of this documentation.

  1. Restart the cinder services

Restart the cinder volume service to allow it to detect the changes made to the cinder.conf file.

  1. Set Metro volumes to ready state

Metro volumes will be set to a Not Ready state after performing rdf pair cleanup. Set these volumes back to Ready state to allow them to be attached to instances. The U4P instance must be restarted for this change to be detected.

# symdev -sid <sid> ready -devs <dev_id1, dev_id2>

Asynchronous and metro replication management groups

Asynchronous and metro volumes in an RDF session, i.e. belonging to an SRDF group, must be managed together for RDF operations (although there is a consistency exempt option for creating and deleting pairs in an Async group). To facilitate this management, we create an internal RDF management storage group on the backend. This RDF management storage group will use the following naming convention:

OS-[rdf_group_label]-[replication_mode]-rdf-sg

It is crucial for correct management that the volumes in this storage group directly correspond to the volumes in the RDF group. For this reason, it is imperative that the RDF group specified in the cinder.conf is for the exclusive use by this Cinder backend. If there are any issues with the state of your RDF enabled volumes prior to performing additional operations in Cinder you will be notified in the Cinder volume logs.

Metro support

SRDF/Metro is a high availability solution. It works by masking both sides of the RDF relationship to the host, and presenting all paths to the host, appearing that they all point to the one device. In order to do this, there needs to be multi-path software running to manage writing to the multiple paths.

Note

The metro issue around formatting volumes when they are added to existing metro RDF groups has been fixed in Unisphere for PowerMax 9.1, however, it has only been addressed on arrays with PowerMax OS and will not be available on arrays running a HyperMax OS.

Volume retype - storage assisted volume migration

Volume retype with storage assisted migration is supported now for PowerMax arrays. Cinder requires that for storage assisted migration, a volume cannot be retyped across backends. For using storage assisted volume retype, follow these steps:

Note

From the Ussuri release of OpenStack the PowerMax driver supports retyping in-use volumes to and from replication enabled volume types with limited exception of volumes with Metro replication enabled. To retype to a volume-type that is Metro enabled the volume must first be detached then retyped. The reason for this is so the paths from the instance to the Metro R1 & R2 volumes must be initialised, this is not possible on the R2 device whilst a volume is attached.

Note

When multiple replication devices are configured. If retyping from one replication mode to another the R1 device ID is preserved and a new R2 side device is created. As a result, the device ID on the R2 array may be different after the retype operation has completed.

Note

Retyping an in-use volume to a metro enabled volume type is not currently supported via storage-assisted migration. This retype can still be performed using host-assisted migration by setting the migration-policy to on-demand.

cinder retype --migration-policy on-demand <volume> <volume-type>
  1. For migrating a volume from one Service Level or Workload combination to another, use volume retype with the migration-policy to on-demand. The target volume type should have the same volume_backend_name configured and should have the desired pool_name to which you are trying to retype to (please refer to Create volume types for details).

    $ cinder retype --migration-policy on-demand <volume> <volume-type>
    

Generic volume group support

Generic volume group operations are performed through the CLI using API version 3.1x of the Cinder API. Generic volume groups are multi-purpose groups which can be used for various features. The PowerMax driver supports consistent group snapshots and replication groups. Consistent group snapshots allows the user to take group snapshots which are consistent based on the group specs. Replication groups allow for tenant facing APIs to enable and disable replication, and to failover and failback, a group of volumes. Generic volume groups have replaced the deprecated consistency groups.

Consistent group snapshot

To create a consistent group snapshot, set a group-spec, having the key consistent_group_snapshot_enabled set to <is> True on the group.

# cinder --os-volume-api-version 3.11 group-type-key GROUP_TYPE set consistent_group_snapshot_enabled="<is> True"

Similarly the same key should be set on any volume type which is specified while creating the group.

# openstack volume type set --property consistent_group_snapshot_enabled="<is> True" POWERMAX_GROUP

If this key is not set on the group-spec or volume type, then the generic volume group will be created/managed by Cinder (not the PowerMax driver).

Note

The consistent group snapshot should not be confused with the PowerMax consistency group which is an SRDF construct.

Replication groups

As with Consistent group snapshot consistent_group_snapshot_enabled should be set to true on the group and the volume type for replication groups. Only Synchronous replication is supported for use with Replication Groups. When a volume is created into a replication group, replication is on by default. The disable_replication api suspends I/O traffic on the devices, but does NOT remove replication for the group. The enable_replication api resumes I/O traffic on the RDF links. The failover_group api allows a group to be failed over and back without failing over the entire host. See below for usage.

Note

A generic volume group can be both consistent group snapshot enabled and consistent group replication enabled.

Storage group names

Storage groups are created on the PowerMax as a result of creation of generic volume groups. These storage groups follow a different naming convention and are of the following format depending upon whether the groups have a name.

TruncatedGroupName_GroupUUID or GroupUUID

Group type, group, and group snapshot operations

Please refer to the official OpenStack block-storage groups documentation for the most up to date group operations

Group replication operations

Generic volume group operations no longer require the user to specify the Cinder CLI version, however, performing generic volume group replication operations still require this setting. When running generic volume group commands set the value --os-volume-api-version to 3.38. These commands are not listed in the latest Cinder CLI documentation so will remain here until added to the latest Cinder CLI version or deprecated from Cinder.

This is how to create a replication group. Please refer to the official OpenStack block-storage groups documentation for the most up to date group operations.

  • Make sure there is a replication_device for Synchronous in cinder.conf

replication_device = backend_id:backend_id_legacy_rep,target_device_id:0001234567890,remote_port_group:PG1,remote_pool:SRP_1,rdf_group_label:os-sync,mode:Synchronous
  • Create a volume type with property replication_enabled=’<is> True’.

$ openstack volume type create --property replication_enabled='<is> True' SYNC_REP_VT
  • Create a Generic group type with extra specs consistent_group_snapshot_enabled=’<is> True’ and consistent_group_replication_enabled=’<is> True’.

$ cinder --os-volume-api-version 3.38 group-type-create GROUP_REP_VT
$ cinder --os-volume-api-version 3.38 group-type-key GROUP_REP_VT set \
  consistent_group_snapshot_enabled='<is> True' \
  consistent_group_replication_enabled='<is> True'
  • Create a Generic group with synchronous volume type SYNC_REP_VT

$ cinder --os-volume-api-version 3.13 group-create --name GROUP_REP GROUP_REP_VT GROUP_REP_VT
  • Create a volume in the Generic group

$ cinder --os-volume-api-version 3.38 create --volume-type SYNC_REP_VT --group-id GROUP_REP \
  --name VOL_REP_GROUP 1
  • Enable group replication

$ cinder --os-volume-api-version 3.38 group-enable-replication GROUP_REP
  • Disable group replication

$ cinder --os-volume-api-version 3.38 group-disable-replication GROUP_REP
  • Failover group

$ cinder --os-volume-api-version 3.38 group-failover-replication GROUP_REP
  • Failback group

$ cinder --os-volume-api-version 3.38 group-failover-replication GROUP_REP \
     --secondary-backend-id default

Manage and unmanage Volumes

Managing volumes in OpenStack is the process whereby a volume which exists on the storage device is imported into OpenStack to be made available for use in the OpenStack environment. For a volume to be valid for managing into OpenStack, the following prerequisites must be met:

  • The volume exists in a Cinder managed pool

  • The volume is not part of a Masking View

  • The volume is not part of an SRDF relationship

  • The volume is configured as a TDEV (thin device)

  • The volume is set to FBA emulation

  • The volume must a whole GB e.g. 5.5GB is not a valid size

  • The volume cannot be a SnapVX target

For a volume to exist in a Cinder managed pool, it must reside in the same Storage Resource Pool (SRP) as the backend which is configured for use in OpenStack. Specifying the pool correctly can be entered manually as it follows the same format:

Pool format: <service_level>+<srp>+<array_id>
Pool example: Diamond+SRP_1+111111111111
Pool values

Key

Value

service_level

The service level of the volume to be managed

srp

The Storage Resource Pool configured for use by the backend

array_id

The PowerMax serial number (12 digit numerical)

Manage volumes

With your pool name defined you can now manage the volume into OpenStack, this is possible with the CLI command cinder manage. The bootable parameter is optional in the command, if the volume to be managed into OpenStack is not bootable leave this parameter out. OpenStack will also determine the size of the value when it is managed so there is no need to specify the volume size.

Command format:

$ cinder manage --name <new_volume_name> --volume-type <powermax_vol_type> \
  --availability-zone <av_zone> <--bootable> <host> <identifier>

Command Example:

$ cinder manage --name powermax_managed_volume --volume-type POWERMAX_ISCSI_DIAMOND \
  --availability-zone nova demo@POWERMAX_ISCSI_DIAMOND#Diamond+SRP_1+111111111111 031D8

After the above command has been run, the volume will be available for use in the same way as any other OpenStack PowerMax volume.

Note

An unmanaged volume with a prefix of OS- in its identifier name cannot be managed into OpenStack, as this is a reserved keyword for managed volumes. If the identifier name has this prefix, an exception will be thrown by the PowerMax driver on a manage operation.

Managing volumes with replication enabled

Whilst it is not possible to manage volumes into OpenStack that are part of a SRDF relationship, it is possible to manage a volume into OpenStack and enable replication at the same time. This is done by having a replication enabled PowerMax volume type (for more information see section Volume Replication) during the manage volume process you specify the replication volume type as the chosen volume type. Once managed, replication will be enabled for that volume.

Note

It is not possible to manage into OpenStack SnapVX linked target volumes, only volumes which are a SnapVX source are permitted. We do not want a scenario where a snapshot source can exist outside of OpenStack management.

Unmanage volume

Unmanaging a volume is not the same as deleting a volume. When a volume is deleted from OpenStack, it is also deleted from the PowerMax at the same time. Unmanaging a volume is the process whereby a volume is removed from OpenStack but it remains for further use on the PowerMax. The volume can also be managed back into OpenStack at a later date using the process discussed in the previous section. Unmanaging volume is carried out using the Cinder unmanage CLI command:

Command format:

$ cinder unmanage <volume_name/volume_id>

Command example:

$ cinder unmanage powermax_test_vol

Once unmanaged from OpenStack, the volume can still be retrieved using its device ID or OpenStack volume ID. Within Unisphere you will also notice that the OS- prefix has been removed, this is another visual indication that the volume is no longer managed by OpenStack.

Manage/unmanage snapshots

Users can manage PowerMax SnapVX snapshots into OpenStack if the source volume already exists in Cinder. Similarly, users will be able to unmanage OpenStack snapshots to remove them from Cinder but keep them on the storage backend.

Set-up, restrictions and requirements:

  1. No additional settings or configuration is required to support this functionality.

  2. Manage/Unmanage snapshots requires SnapVX functionality support on PowerMax.

  3. Manage/Unmanage Snapshots in OpenStack Cinder is only supported at present through Cinder CLI commands.

  4. It is only possible to manage or unmanage one snapshot at a time in Cinder.

Manage SnapVX snapshot

It is possible to manage PowerMax SnapVX snapshots into OpenStack, where the source volume from which the snapshot is taken already exists in, and is managed by OpenStack Cinder. The source volume may have been created in OpenStack Cinder, or it may have been managed in to OpenStack Cinder also. With the support of managing SnapVX snapshots included in OpenStack Queens, the restriction around managing SnapVX source volumes has been removed.

Note

It is not possible to manage into OpenStack SnapVX linked target volumes, only volumes which are a SnapVX source are permitted. We do not want a scenario where a snapshot source can exist outside of OpenStack management.

Requirements/restrictions:

  1. The SnapVX source volume must be present in and managed by Cinder.

  2. The SnapVX snapshot name must not begin with OS-.

  3. The SnapVX snapshot source volume must not be in a failed-over state.

  4. Managing a SnapVX snapshot will only be allowed if the snapshot has no linked target volumes.

Command structure:

  1. Identify your SnapVX snapshot for management on the PowerMax, note the name.

  2. Ensure the source volume is already managed into OpenStack Cinder, note the device ID.

  3. Using the Cinder CLI, use the following command structure to manage a Snapshot into OpenStack Cinder:

$ cinder snapshot-manage --id-type source-name
                         [--name <name>]
                         [--description <description>]
                         [--metadata [<key=value> [<key=value> ...]]]
                         <volume name/id> <identifier>

Positional arguments:

  • <volume name/id> Source OpenStack volume name

  • <identifier> Name of existing snapshot on PowerMax backend

Optional arguments:

  • --name <name> Snapshot name (Default=``None``)

  • --description <description> Snapshot description (Default=``None``)

  • --metadata [<key=value> [<key=value> ...]] Metadata key=value pairs (Default=``None``)

Example:

$ cinder snapshot-manage --name SnapshotManaged \
                         --description "Managed Queens Feb18" \
                         powermax-vol-1 PowerMaxSnapshot

Where:

  • The name in OpenStack after managing the SnapVX snapshot will be SnapshotManaged.

  • The snapshot will have the description Managed Queens Feb18.

  • The Cinder volume name is powermax-vol-1.

  • The name of the SnapVX snapshot on the PowerMax backend is PowerMaxSnapshot.

Outcome:

After the process of managing the Snapshot has completed, the SnapVX snapshot on the PowerMax backend will be prefixed by the letters OS-, leaving the snapshot in this example named OS-PowerMaxSnapshot. The associated snapshot managed by Cinder will be present for use under the name SnapshotManaged.

Unmanage cinder snapshot

Unmanaging a snapshot in Cinder is the process whereby the snapshot is removed from and no longer managed by Cinder, but it still exists on the storage backend. Unmanaging a SnapVX snapshot in OpenStack Cinder follows this behaviour, whereby after unmanaging a PowerMax SnapVX snapshot from Cinder, the snapshot is removed from OpenStack but is still present for use on the PowerMax backend.

Requirements/Restrictions:

  • The SnapVX source volume must not be in a failed over state.

Command Structure:

Identify the SnapVX snapshot you want to unmanage from OpenStack Cinder, note the snapshot name or ID as specified by Cinder. Using the Cinder CLI use the following command structure to unmanage the SnapVX snapshot from Cinder:

$ cinder snapshot-unmanage <snapshot>

Positional arguments:

  • <snapshot> Cinder snapshot name or ID.

Example:

$ cinder snapshot-unmanage SnapshotManaged

Where:

  • The SnapVX snapshot name in OpenStack Cinder is SnapshotManaged.

After the process of unmanaging the SnapVX snapshot in Cinder, the snapshot on the PowerMax backend will have the OS- prefix removed to indicate it is no longer OpenStack managed. In the example above, the snapshot after unmanaging from OpenStack will be named PowerMaxSnapshot on the storage backend.

List manageable volumes and snapshots

Manageable volumes

Volumes that can be managed by and imported into Openstack.

List manageable volume is filtered by:

  • Volume size should be 1026MB or greater (1GB PowerMax Cinder Vol = 1026 MB)

  • Volume size should be a whole integer GB capacity

  • Volume should not be a part of masking view.

  • Volume status should be Ready

  • Volume service state should be Normal

  • Volume emulation type should be FBA

  • Volume configuration should be TDEV

  • Volume should not be a system resource.

  • Volume should not be private

  • Volume should not be encapsulated

  • Volume should not be reserved

  • Volume should not be a part of an RDF session

  • Volume should not be a SnapVX Target

  • Volume identifier should not begin with OS-.

  • Volume should not be in more than one storage group.

Manageable snaphots

Snapshots that can be managed by and imported into Openstack

List manageable snapshots is filtered by:

  • The source volume should be marked as SnapVX source.

  • The source volume should be 1026MB or greater

  • The source volume should be a whole integer GB capacity.

  • The source volume emulation type should be FBA.

  • The source volume configuration should be TDEV.

  • The source volume should not be private.

  • The source volume should be not be a system resource.

  • The snapshot identifier should not start with OS- or temp-.

  • The snapshot should not be expired.

  • The snapshot generation number should npt be greater than 0.

Note

There is some delay in the syncing of the Unisphere for PowerMax database when the state/properties of a volume is modified using symcli. To prevent this it is preferable to modify state/properties of volumes within Unisphere.

Cinder backup support

PowerMax Cinder driver support Cinder backup functionality. For further information on setup, configuration and usage please see the official OpenStack volume backup documentation and related volume backup CLI guide.

Note

rpc_response_timeout may need to be increased significantly in volume backup operations especially in replication scenarios where the creation operation will be longer. For more information on rpc_response_timeout please refer to Live migration configuration

Port group & port load balancing

By default port groups are selected at random from cinder.conf when connections are initialised between volumes on the backend array and compute instances in Nova. If a port group is set in the volume type extra specifications this will take precedence over any port groups configured in cinder.conf. Port selection within the chosen port group is also selected at random by default.

With port group and port load balancing in the PowerMax for Cinder driver users can now select the port group and port load by determining which has the lowest load. The load metric is defined by the user in both instances so the selection process can better match the needs of the user and their environment. Available metrics are detailed in the performance metrics section.

Port Groups are reported on at five minute time deltas (diagnostic), and FE Ports are reported on at one minute time deltas (real-time) if real-time metrics are enabled, else default five minute time delta (diagnostic). The window at which performance metrics are analysed is a user-configured option in cinder.conf, this is detailed in the configuration section.

Calculating load

The process by which Port Group or Port load is calculated is the same for both. The user specifies the look back window which determines how many performance intervals to measure, 60 minutes will give 12 intervals of 5 minutes each for example. If no lookback window is specified or is set to 0 only the most recent performance metric will be analysed. This will give a slight performance improvement but with the improvements made to the performance REST endpoints for load this improvement is negligible. For real-time stats a minimum of 1 minute is required.

Once a call is made to the performance REST endpoints, the performance data for that PG or port is extracted. Then the metric values are summed and divided by the count of intervals to get the average for the look back window.

The performance metric average value for each asset is added to a Python heap. Once all assets have been measured the lowest value will always be at position 0 in the heap so there is no extra time penalty requirement for search.

Pre-requisites

Before load balancing can be enabled in the PowerMax for Cinder driver performance metrics collection must be enabled in Unisphere. Real-time performance metrics collection is enabled separately from diagnostic metrics collection. Performance metric collection is only available for local arrays in Unisphere.

After performance metrics registration there is a time delay before Unisphere records performance metrics, adequate time must be given before enabling load balancing in Cinder else default random selection method will be used. It is recommended to wait 4 hours after performance registration before enabling load balancing in Cinder.

Configuration

A number of configuration options are available for users so load balancing can be set to better suit the needs of the environment. These configuration options are detailed in the table below.

Load balance cinder.conf configuration options

cinder.conf parameter

options

Default

Description

load_balance

True/False

False

Enable/disable load balancing for
a PowerMax backend.

load_balance_real_time

True/False

False

Enable/disable real-time performance
metrics for Port level metrics
(not available for Port Group).

load_data_format

Avg/Max

Avg

Performance data format, not
applicable for real-time.

load_lookback

int

60

How far in minutes to look back for
diagnostic performance metrics in
load calculation, minimum of 0
maximum of 1440 (24 hours).

load_real_time_lookback

int

1

How far in minutes to look back for
real-time performance metrics in
load calculation, minimum of 1
maximum of 60 (24 hours).

port_group_load_metric

See below

PercentBusy

Metric used for port group load
calculation.

port_load_metric

See below

PercentBusy

Metric used for port group load
calculation.

Port-Group Metrics

Port-group performance metrics

Metric

cinder.conf option

Description

% Busy

PercentBusy

The percent of time the port group is busy.

Avg IO Size (KB)

AvgIOSize

Calculated value: (HA Kbytes transferred per sec /
total IOs per sec)

Host IOs/sec

IOs

The number of host IO operations performed each second,
including writes and random and sequential reads.

Host MBs/sec

MBs

The number of host MBs read each second.

MBs Read/sec

MBRead

The number of reads per second in MBs.

MBs Written/sec

MBWritten

The number of writes per second in MBs.

Reads/sec

Reads

The average number of host reads performed per second.

Writes/sec

Writes

The average number of host writes performed per second.

Port Metrics

Port performance metrics

Metric

cinder.conf option

Real-Time Supported

Description

% Busy

PercentBusy

Yes

The percent of time the port is busy.

Avg IO Size (KB)

AvgIOSize

Yes

Calculated value: (HA Kbytes transferred per sec /
total IOs per sec)

Host IOs/sec

IOs

Yes

The number of host IO operations performed each second,
including writes and random and sequential reads.

Host MBs/sec

MBs

Yes

The number of host MBs read each second.

MBs Read/sec

MBRead

Yes

The number of reads per second in MBs.

MBs Written/sec

MBWritten

Yes

The number of writes per second in MBs.

Reads/sec

Reads

Yes

The number of read operations performed by the port per second.

Writes/sec

Writes

Yes

The number of write operations performed each second by the port.

Speed Gb/sec

SpeedGBs

No

Speed.

Response Time (ms)

ResponseTime

No

The average response time for the reads and writes.

Read RT (ms)

ReadResponseTime

No

The average time it takes to serve one read IO.

Write RT (ms)

WriteResponseTime

No

The average time it takes to serve one write IO.

Upgrading from SMI-S based driver to REST API based driver

Seamless upgrades from an SMI-S based driver to REST API based driver, following the setup instructions above, are supported with a few exceptions:

  1. Seamless upgrade from SMI-S(Ocata and earlier) to REST(Pike and later) is now available on all functionality including Live Migration.

  2. Consistency groups are deprecated in Pike. Generic Volume Groups are supported from Pike onwards.

Known issues

These known issues exist in the current release of OpenStack: