VMware

vSAN node rollback

Anyone have the procedure to roll back a vSAN node.
i.e an update from 6.5 -> 6.7 on remediation something has gone sideways where access via ssh has died hard reboot isnt helping etc. node was fully evac before remediation.

looks like /etc/vmware/esx.conf.lock timeout waiting 30 seconds
lock is currently held by process xyz /sbin/sh /sbin/backup.sh

Is it safe to try the alt-bootbank on a vSAN node to determine if its a upgrade issue or SD card issue?


View Reddit by badabing888View Source

Related Articles

2 Comments

  1. Yes, the alt boot bank works fine on vSAN nodes (we’ve used this with bad driver/firmware combinations that caused us grief); the on disk format doesn’t change during the upgrade, so vSAN will just figure itself out. Good luck!

  2. The alt-boot bank should work. However if there has been a second reboot it may have been overridden. This can be confirmed prior to trying via cli.
    If it has, next best thing is to perform a vib install of your old bundle.

    If this is the last host in the cluster, depending on the 6.5 version, replication mode may have been changed from multicast to unicast. Im unsure of the implications if you revert back. It’s only applicable for 6.5GA (maybe u1) but unsure what version you’re using.

    Edit:
    tail -2 /*bootbank/boot.cfg will show you both active and alt bootbank versions.

    The cluster will automatically switch back from unicast to multicast if a host earlier than 6.5e is introduced.

Leave a Reply

Close