Hi,

An ext3 fs on my system has journaled quota turned on via the "usrjquota" mount option, and "quotacheck" is invoked by init script during boot time.
As the number of files grows, quotacheck takes longer and longer to finish.
In my understanding, journaled quota can survive system crash or other un-clean shutdown. Running quotacheck seems to be redundant.

I'd like to know : what is the usual (and safe) practice of using quotacheck on fs with journaled quota ?

"Journaling" refers to a process of recording redundant data that is specifically designed to expedite filesystem recovery. It is, literally, a "journal" of recently-performed operations.

"Quota checking" is unrelated to journaling, and, yes, it can be expensive, because current resource-usage must generally be calculated (and, re-calculated), to determine if the quota limit has been exceeded. I find it to be of limited usefulness.

Do not be tempted to turn off journaling. In my experience (and, yes, I have some ... ), a filesystem is pragmatically un-recoverable without it.

As far as I know, quotacheck only really needs to be run if the filesystem is marked as unclean for non-journalled quotas.

However, as you're using journalled quotas, these have the advantage that no quota check is required, even after an unclean shutdown.

So, no need for quotacheck.

1 members found this post helpful.