Details
-
Bug
-
Status: Done
-
Medium
-
Resolution: Fixed
-
2.33.0
-
None
-
I've tested PMM OVA on virsh/KVM.
pmm2-server-2.30.0.ova
pmm-server-2.33.0.ova
-
1
-
Yes
-
Yes
-
No
-
C/S Core
Description
Description :
Copy of resize-xfs-lvm to resize-xfs failed during PMM upgrade on a non-Docker installation.
Error from /srv/logs/pmm-update-perform.log
TASK [Copy rezise-xfs file for lvm] ********************************************
fatal: [localhost]: FAILED! => {"changed": false, "msg": "file (/var/lib/cloud/scripts/per-boot/resize-xfs) is absent, cannot continue", "path": "/var/lib/cloud
/scripts/per-boot/resize-xfs", "state": "absent"}
[WARNING]: The src option requires state to be 'link' or 'hard'. This will
become an error in Ansible 2.10PLAY RECAP *********************************************************************
localhost : ok=68 changed=15 unreachable=0 failed=1 skipped=25 rescued=0 ignored=0 time="2022-12-19T00:21:56Z" level=fatal msg="RunPlaybook failed: exit status 2"
ProjectName: pmm-update
Version: 2.33.0
How to reproduce :
1) Remove/rename existing resize-xfs file or it can also be reproduce if an upgrade is done on PMM 2.24 and below as resize-xfs is still not present on such version.
cd /var/lib/cloud/scripts/per-boot/ mv -nv resize-xfs resize-xfs.bkp
2) Perform the upgrade :
supervisorctl start pmm-update-perform
3) The error will be encountered and the upgrade will automatically restart in a continuous loop.
Workaround :
Search for the file :
# ls -lh $(find / -iname 'resize-xfs*' -print 2>/dev/null) -rw-r--r--. 1 root root 642 Nov 29 06:17 /usr/share/pmm-update/ansible/playbook/tasks/files/resize-xfs-lvn -rwxr-xr-x. 1 root root 642 Dec 6 18:15 /var/lib/cloud/scripts/per-boot/resize-xfs.bkp
Manually copy resize-xfs-lvn to resize-xfs :
cp -nv /usr/share/pmm-update/ansible/playbook/tasks/files/resize-xfs-lvn /var/lib/cloud/scripts/per-boot/resize-xfs
3) The upgrade will now succeed, but if you rebooted the machine you have to re-run :
supervisorctl start pmm-update-perform
4) Check log and you should see like below with failed=0 :
# tail /srv/logs/pmm-update-perform.log ] } TASK [Delete content & directory] ********************************************** changed: [localhost] PLAY RECAP ********************************************************************* localhost : ok=85 changed=19 unreachable=0 failed=0 skipped=25 rescued=0 ignored=0 time="2022-12-19T00:24:27Z" level=info msg="Waiting for Grafana dashboards update to finish..."
Suggested fix :
As discussed in Slack thread #pmm https://perconacorp.slack.com/archives/C03J8FZFU/p1670818275943949
1 ) Edit /usr/share/pmm-update/ansible/playbook/tasks/update.yml. Use copy instead of file. Addionally, please correct task name: rezise-xfs to resize-xfs
From :
# https://jira.percona.com/browse/PMM-9298 - name: Copy rezise-xfs file for lvm file: src: resize-xfs-lvm dest: /var/lib/cloud/scripts/per-boot/resize-xfs mode: 0755 when: not is_docker
To :
# https://jira.percona.com/browse/PMM-9298 - name: Copy resize-xfs file for lvm copy: src: resize-xfs-lvm dest: /var/lib/cloud/scripts/per-boot/resize-xfs mode: 0755 when: not is_docker
2 ) Rename resize-xfs-lvn to resize-xfs-lvm as it the one declared as the source
# cd /usr/share/pmm-update/ansible/playbook/tasks/files/ # mv -nv resize-xfs-lvn resize-xfs-lvm