Do I need to run fsck?
How often should I run it?
Please shutdown the apps, run autoconfig and restart the apps.
If you restart the app tier server, the OS should force a fsck if it has been too long. Running fsck manually is not a task for DBAs who do not want to do filesystem recoveries. I may or may not be speaking from experience here.
Is it a symptom of some disk corruption or failure?
Does fsck run in a VM environment? Is it the same fsck being run on the host server of the VM?
I am confused of running fsck between the VM and the Host server (baremetal) itself
or if there is such a commands and what are the difference of the two fsck?
I don't think this type of problem can be fixed by fsck but rather autoconfig should fix it.
fsck is for fixing corrupt file system but if the corruption is severe, then it may fix the file references on the inode level but the content of the actual file will be a mess/corrupted. You will need a backup of that file in such a case.
Usually, if the container level filesystem is corrupted, then you won't be even able to start your VM . Also, don't run fsck in multi-user environments. If you need to run fsck, boot into single user mode and then run fsck. Most Linux OSes will check whether they need to run fsck automatically at startup so you should normally not need to worry about fsck.
Thanks Handat and Michael,
I just can not explain to my boss when she ask why suddenly the file is missing? or can not be read?
I can only guess that it is a symptoms of disk corruption? which is logical.
Or does this message PARAMETER.CONFIG missing is just a warning only and not really the primary cause of the issue?
And I can disregard it? or maybe doing a restart of apps and db tiers will resolve the issue?
A file corruption could result in the file gone missing, but it could also be the case that someone deleted it or it never existed.
Checking your backups for the missing file might shed some light on whether it existed before.