VCSA and fsck after power issue.

[hendersp3]

Can anyone comment on why fsck fails and needs to be run manually?  I am curious.  about what specifically requires a manual fsck run.  I have an environment where unfortunately power is less than stable and the vcsa, log insight and nsx appliances (all photon) come back with file system errors after booting from a power issue.  This always happens, every time.  Our engineering team has tried countless different scenarios to reproduce this in an effort to avoid or limit the need to manually run fsck but actually have struggled to make fsck fail.

Funny we cannot break it as hard as we try but in one particular installation it breaks every time.  Any ideas of what might make a difference?  I am not asking how to run fsck and repair it.  I am trying to solve for \understand why it happens.

Thanks for any ideas or suggestions. 

[msripada]

fsck can fail at any stage, it can be due to some file were not properly written as most of the latest filesystems including photon os is journaling filesystems. If the filesystem is attempting to write and could not complete it then the journal would have been corrupted.

Most of the VCSA partitions are lvms, if the lvms did not mount properly by the time fsck started to correct the corruptions can also lead to this.

In general, linux machines are sensitive to storage/power outages. When you fail with fsck, run the journalctl which can give you more info on which partition it failed. May be by tracking that partition, we may try to track back the issue what would have happened on that disk

Thanks,

MS