Bug 1226267

Summary: Update of Azure python SDK to 3.11 breaks Azure Fence Agent
Product: [openSUSE] PUBLIC SUSE Linux Enterprise High Availability Extension 15 SP5 Reporter: Dennis Padia <dennis.padia>
Component: FencingAssignee: SUSE Linux Enterprise High Availability Team <ha-bugs>
Status: CONFIRMED --- QA Contact: Michele Pagot <michele.pagot>
Severity: Critical    
Priority: P5 - None CC: amach, hp.jansen, hsehic, Ralitza.Deltcheva
Version: unspecified   
Target Milestone: ---   
Hardware: x86-64   
OS: Other   
Whiteboard:
Found By: --- Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---

Description Dennis Padia 2024-06-12 21:13:32 UTC 
We have an issue with Azure Fence Agent in SLES 15 SP4 and SP5. When we update the SLES 15 SP4 and SP5 with “zypper update” command, it updates Azure python SDK from python3-azure-sdk to python311-azure-sdk. 

With the introduction of new Python 3.11 interpreter (https://www.suse.com/c/incompatible-changes-ahead-for-public-cloud-sdks/), fence_azure_arm is failing with error message “Azure Resource Manager Python SDK not found or not accessible”. 

Azure fence agent is not updated to handle the updated Azure python SDK 3.11. So, could you please help on this issue? 

Below are the details on the issue or on how to reproduce

1)	Created a new VM with SLES 15 SP5, it has below set of package versions. The Azure fence agent resource works fine. 

testvm1:~ # rpm -qa | grep -i python3-azure-mgmt-compute
python3-azure-mgmt-compute-26.1.0-150200.9.3.2.noarch

testvm1:~ # rpm -qa | grep -i python3-azure-identity
python3-azure-identity-1.10.0.0-150200.6.3.2.noarch

testvm1:~ # more /etc/os-release
NAME="SLES"
VERSION="15-SP5"
VERSION_ID="15.5"
PRETTY_NAME="SUSE Linux Enterprise Server 15 SP5"
ID="sles"
ID_LIKE="suse"
ANSI_COLOR="0;32"
CPE_NAME="cpe:/o:suse:sles:15:sp5"
VARIANT_ID="sles-sap"

testvm1:~ # rpm -qa | grep -i resource-agents
resource-agents-4.12.0+git30.7fd7c8fa-150500.3.6.2.x86_64

testvm1:~ # rpm -qa | grep -i fence-agents
fence-agents-4.12.1+git.1677142927.bf55c675-150500.4.6.1.x86_64

testvm1:~ # fence_azure_arm --msi --action=list --resourceGroup=dp-sles --subscriptionId=e663cc2d-722b-4be1-b636-bbd9e4c60fd9
testvm1,
testvm2,

testvm1:~ # crm status
Status of pacemakerd: 'Pacemaker is running' (last updated 2024-06-12 19:28:11Z)
Cluster Summary:
  * Stack: corosync
  * Current DC: testvm1 (version 2.1.5+20221208.a3f44794f-150500.6.14.4-2.1.5+20221208.a3f44794f) - partition with quorum
  * Last updated: Wed Jun 12 19:28:12 2024
  * Last change:  Wed Jun 12 19:26:11 2024 by root via cibadmin on testvm1
  * 2 nodes configured
  * 1 resource instance configured

Node List:
  * Online: [ testvm1 testvm2 ]

Full List of Resources:
  * rsc_st_azure        (stonith:fence_azure_arm):       Started testvm1

2)	UPDATE OS: Update OS with “zypper update” command, which update Azure SDK to 3.11. The “fence_azure_arm” command fails with error message “Azure Resource Manager Python SDK not found or not accessible”

testvm1:~ # rpm -qa | grep -i python3-azure-mgmt-compute
testvm1:~ # rpm -qa | grep -i python3-azure-identity

testvm1:~ # rpm -qa | grep -i python311-azure-mgmt-compute
python311-azure-mgmt-compute-30.5.0-150400.14.3.1.noarch

testvm1:~ # rpm -qa | grep -i python311-azure-identity
python311-azure-identity-1.15.0-150400.11.3.1.noarch
python311-azure-identity-broker-1.0.0-150400.9.3.1.noarch

testvm1:~ # fence_azure_arm --msi --action=list --resourceGroup=dp-sles --subscriptionId=e663cc2d-722b-4be1-b636-bbd9e4c60fd9
2024-06-12 19:40:05,371 ERROR: Azure Resource Manager Python SDK not found or not accessible

2024-06-12 19:40:05,371 ERROR: Please use '-h' for usage

testvm1:~ # crm status
Status of pacemakerd: 'Pacemaker is running' (last updated 2024-06-12 21:03:13Z)
Cluster Summary:
  * Stack: corosync
  * Current DC: testvm2 (version 2.1.5+20221208.a3f44794f-150500.6.14.4-2.1.5+20221208.a3f44794f) - partition with quorum
  * Last updated: Wed Jun 12 21:03:14 2024
  * Last change:  Wed Jun 12 20:21:30 2024 by hacluster via crmd on testvm1
  * 2 nodes configured
  * 1 resource instance configured

Node List:
  * Online: [ testvm1 testvm2 ]

Full List of Resources:
  * rsc_st_azure        (stonith:fence_azure_arm):       Stopped

Failed Resource Actions:
  * rsc_st_azure_start_0 on testvm2 'error' (1): call=20, status='complete', last-rc-change='Wed Jun 12 20:21:34 2024', queued=0ms, exec=3221ms
  * rsc_st_azure_start_0 on testvm1 'error' (1): call=20, status='complete', last-rc-change='Wed Jun 12 20:21:30 2024', queued=0ms, exec=3227ms