subcloud1 and subcloud2 device label get removed after "dcmanager fw-update-strategy apply"

[Difu Hu]

Brief Description
-----------------
subcloud1 and subcloud2 device label get removed.
subcloud3 and subcloud4 device label are still there.

Severity
--------
Major

Steps to Reproduce
------------------
dcmanager fw-update-strategy create
dcmanager fw-update-strategy apply

Expected Behavior
------------------
all subcloud keep device label after FPGA image update

Actual Behavior
----------------
subcloud1 and subcloud2 device label get removed

Reproducibility
---------------
permanent

System Configuration
--------------------
Lab-name: DC-3

Branch/Pull Time/Commit
-----------------------
2020-07-31_20-00-00

Last Pass
---------
N/A

Timestamp/Logs
--------------
subcloud1:
[sysadmin@controller-0 ~(keystone_admin)]$ system device-label-list

[sysadmin@controller-0 ~(keystone_admin)]$ system device-image-list
+--------------------------------------+----------------+------------+------------+--------------+---------------+---------------+------+-------------+---------------+---------+----------------+
| uuid | bitstream_type | pci_vendor | pci_device | bitstream_id | key_signature | revoke_key_id | name | description | image_version | applied | applied_labels |
+--------------------------------------+----------------+------------+------------+--------------+---------------+---------------+------+-------------+---------------+---------+----------------+
| 4868ef33-bbe3-44a0-8d8b-a0454a726bfa | functional | 8086 | 0b30 | 1 | None | None | None | None | None | False | None |
+--------------------------------------+----------------+------------+------------+--------------+---------------+---------------+------+-------------+---------------+---------+----------------+

subcloud3:
[sysadmin@controller-0 ~(keystone_admin)]$ system device-label-list
+--------------+------------------+-----------+-------------+
| hostname | PCI device name | label key | label value |
+--------------+------------------+-----------+-------------+
| controller-0 | pci_0000_b4_00_0 | subcloud | abc |
+--------------+------------------+-----------+-------------+
[sysadmin@controller-0 ~(keystone_admin)]$ system device-image-list
+--------------------------------------+----------------+------------+------------+--------------+---------------+---------------+------+-------------+---------------+---------+-------------------------+
| uuid | bitstream_type | pci_vendor | pci_device | bitstream_id | key_signature | revoke_key_id | name | description | image_version | applied | applied_labels |
+--------------------------------------+----------------+------------+------------+--------------+---------------+---------------+------+-------------+---------------+---------+-------------------------+
| 4868ef33-bbe3-44a0-8d8b-a0454a726bfa | functional | 8086 | 0b30 | 1 | None | None | None | None | None | True | [{u'subcloud': u'abc'}] |
+--------------------------------------+----------------+------------+------------+--------------+---------------+---------------+------+-------------+---------------+---------+-------------------------+
[sysadmin@controller-0 ~(keystone_admin)]$ system device-image-state-list
+--------------+----------+--------------------------------------+-----------+---------------+----------------------------------+
| hostname | PCI | Device image uuid | status | Update start | updated_at |
| | device | | | time | |
| | address | | | | |
+--------------+----------+--------------------------------------+-----------+---------------+----------------------------------+
| controller-0 | 0000:b4: | 4868ef33-bbe3-44a0-8d8b-a0454a726bfa | completed | 2020-08-03T17 | 2020-08-03T17:56:43.110604+00:00 |
| | 00.0 | | | :23:49. | |
| | | | | 847639+00:00 | |
| | | | | | |
+--------------+----------+--------------------------------------+-----------+---------------+----------------------------------+

Test Activity
-------------
Functional Testing

[Ghada Khalil (gkhalil)]

As per Chris Friesen, this is expected to be addressed by this in-progress review:
https://review.opendev.org/#/c/744684/

Assigning to Teresa Ho to confirm and update the LP once the fix is merged.

[Ghada Khalil (gkhalil)]

Marking as stx.5.0 - issue related to DC FPGA Orchestration feature

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/744684
Committed: https://git.openstack.org/cgit/starlingx/config/commit/?id=046f98f93b50338c5e392406b82093c5f33e2fca
Submitter: Zuul
Branch: master

commit 046f98f93b50338c5e392406b82093c5f33e2fca
Author: Teresa Ho <email address hidden>
Date: Tue Aug 4 00:10:24 2020 -0400

    sysinv: misc fixes for firmware update

    After reboot, the pci address of the N3000 device may change and new
    pci_device DB entries are created by sysinv-agent to replace the old
    pci devices created by sysinv-fpga-agent.
    This causes all the dependents (device_label and device_image_state)
    of the pci_device records to disappear.
    This problem is solved by only having sysinv-fpga-agent remove N3000
    device entries.

    This commit also sets the host device_image_update attribute to
    pending when all device updates operations are processed and
    any of the updates failed.

    The applied-labels shown in the device-image-list shows duplicate
    name-value pair if multiple devices have the same label. This update
    removes the duplicates in the list.
    The comments for the applied field is also updated.

    Renamed the RPC function remove_device_image() to
    clear_device_image_alarm().

    Story: 2007875
    Task: 40392
    Closes-Bug: 1890296

    Change-Id: I8e76e13d0a186c21531c8ca4eca398149d6b7db4
    Signed-off-by: Teresa Ho <email address hidden>

[Difu Hu (difuhu)]

Verified on build 2020-08-06_20-00-00.