oVirt Node boots into ermergency mode after adding in Engine #88
Comments
|
Thank you for the detailed bug report! There are a few root causes I could think of, either related to the setup (3.3. Preparing FCP Storage) or the changes that were introduced with a switch from el8 to el9. Before I get into the guessing game, maybe @didib, @nirs, or @lveyde know more already? Regarding "The failure is LVM on several /var Volumes" -> The default |
|
Hello, looks like all /var Volumes are missing:
[root@ovnode02 ~]# df -Th journalctl -xb -- Logs begin at Fri 2022-10-14 15:45:30 CEST, end at Fri 2022-10-14 15:53:36 CEST. -- |
Nothing that comes to my mind right away. |
|
Thanks @lveyde, but I'm working on other projects now, so you should This looks like the known issue with ovirt node upgrade. When ovirt-node runs To fix this you need to remove the the devices file: and reboot. After reboot, run again: To create a proper configuration. |
|
Great, thanks @nirs ! |
|
Hello, I deleted /etc/lvm/devices/system.devices after adding the Node to the Engine in emergency mode and rebooted, the node came back in emergency mode with the same errors. |
Have you tried to run vdsm-tool config-lvm-filter ? |
|
Yes, thats what I tried until now: deleted /etc/lvm/devices/system.devices -> reboot -> ermegency mode -> vdsm-tool config-lvm-filter -> reboot -> ermergency mode -> deleted /etc/lvm/devices/system.devices + vdsm-tool config-lvm-filter -> reboot -> ermergency mode |
|
@humnab You need to activate the gluster volumes before running The LVM devices that you see running These are the detailed steps to fix it: |
|
@aesteve-rh I have done: [root@ovnode01 ~]# rm /etc/lvm/devices/system.devices logical volume: /dev/mapper/onn_ovnode01-home logical volume: /dev/mapper/onn_ovnode01-ovirt--node--ng--4.5.2--0.20220810.0+1 logical volume: /dev/mapper/onn_ovnode01-swap logical volume: /dev/mapper/onn_ovnode01-tmp logical volume: /dev/mapper/onn_ovnode01-var logical volume: /dev/mapper/onn_ovnode01-var_crash logical volume: /dev/mapper/onn_ovnode01-var_log logical volume: /dev/mapper/onn_ovnode01-var_log_audit logical volume: /dev/mapper/onn_ovnode01-var_tmp Configuring LVM system.devices. onn_ovnode01 Configure host? [yes,NO] yes Please reboot to verify the configuration. |
|
After the reboot, the host booted to emergency mode again:
|
|
@humnab could you please paste the output of the lsblk command? |
|
This is after vdsm-tool config-lvm-filter + reboot: |
|
And this is after activating the volumes (no gluster volumes, all is FC block): [root@ovnode01 ~]# rm /etc/lvm/devices/system.devices |
|
There are 2 pvs (sda, sdb) under the same vg (onn_ovnode01), which is imported in the lvm devices file. That looks sane to me. The devices files should be correct and the device visible to lvm, there is something else preventing the volumes to be mounted during boot, it seems. |
|
sdc + sdd should be also onn_ovnode01 (4 Paths, all active, 2 preferred) [root@ovnode01 ~]# multipath -ll I don't think that the Storage is the problem, on the same Storage Type (IBM SVC) a Intel Node wit 8 Gbit FC HBAs works...(also with a 1,2 TB disk) |
Describe the bug
To reproduce
Expected behavior
Screenshots
Additional context
tried several options:
ovirt-node-ng-installer-4.5.2-2022081013.el8.iso -> does not boot after adding
ovirt-node-ng-installer-4.5.2-2022081013.el9.iso -> does boot after adding
Centos Stream 8 -> does boot after adding
Rocky Linux 8.6 -> does boot after adding
The Server is a Supermicro 1024US-TRT, 2 Sockets, 2TB Ram
Boot is from FC SAN! HBA: Qlogic qle2672
Storage is a IBM Storwize Fabric attached (also tried a old DS3500 direct and Fabric attached, same result)
Boot from a internal Disk works!
I'm a little bit lost in which logs I should provide? Thanks!
The text was updated successfully, but these errors were encountered: