IBM SVF driver: Detaching from second instance fails for multi-attach volumes

[saikrishnagk]

Description : During detach operation for the Multi attach type volumes the first detach from the host is successful. In the second detach operation the volume remains in detaching state.

Reason for failure : After the first detach the host-related info is getting deleted in the terminate_connection function. while detaching from the second instance the volume remains in detaching state with host_name not found error.

Steps to reproduce the issue

1. Create a multi attached volume

2. Spin up two instances

3. Attach the volume to both instances

4. Detach the volume from the first instance

    Detached Successfully

5. Detach the volume from the second instance

     Volume remains in detaching state after few minutes

+--------------------------------------+-----------+------------------+------+-----------------+----------+--------------------------------------+
| ID | Status | Name | Size | Volume Type | Bootable | Attached to |
+--------------------------------------+-----------+------------------+------+-----------------+----------+--------------------------------------+
| 26eaddf8-b198-4a16-aaf7-88b66a7a9d13 | detaching | gm_vol_auto_0 | 1 | vol_type_auto_2 | false | afec06c8-1309-4401-ac90-4c4c42646a52 |

Cinder Log:

Aug 17 11:30:10 devstack cinder-volume[311264]: ERROR oslo_messaging.rpc.server File "/opt/stack/cinder/cinder/volume/manager.py", line 4833, in _connection_terminate
Aug 17 11:30:10 devstack cinder-volume[311264]: ERROR oslo_messaging.rpc.server raise exception.VolumeBackendAPIException(data=err_msg)
Aug 17 11:30:10 devstack cinder-volume[311264]: ERROR oslo_messaging.rpc.server cinder.exception.VolumeBackendAPIException: Bad or unexpected response from the storage volume backend API: Terminate volume connection failed: Volume driver reported an error: Volume driver reported an error: _get_map_info_from_connector: Failed to get host name from connector.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.opendev.org/c/openstack/cinder/+/809518

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/c/openstack/cinder/+/809518
Committed: https://opendev.org/openstack/cinder/commit/c569b7f22575ef22f9d0d2af512b31c27ca0ccbe
Submitter: "Zuul (22348)"
Branch: master

commit c569b7f22575ef22f9d0d2af512b31c27ca0ccbe
Author: gksk <email address hidden>
Date: Fri Sep 17 04:18:18 2021 +0000

    [SVf] Fixed Detach for multi-attach volumes

    [Spectrum Virtualize family] During detach operation
    for Multi attach type volumes the first detach
    from the host is successful. In the second detach
    operation the volume remains in detaching state.

    Fixed the above issue by adding a check for multi-attach
    volumes and get the host mapping info of a volume.
    if the volume is attached to single host skip the
    terminate connection function and return.

    closes bug: #1941694

    Change-Id: I40c088c17b15b3de815caddfe2922097228e3b00

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/cinder 20.0.0.0rc1 release candidate.