From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.1) Gecko/20021003
Description of problem:
fsck -T -a $fsckoptions /
A better way is
fsck -T -a / $fsckoptions
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. turn your machine on
2. turn it off many times without a soft shutdown
3. ext3 will need a manual fsck
Actual Results: 3. ext3 will need a manual fsck
Expected Results: We would like to never make a human fsck.
We are using Red Hat Linux for 1700 POS (point of sale) accross the country,
being used by unskilled users that like to unsoftly shut the machine down. After
some shutdowns, rc.sysinit will ask for a manual fsck. This is unacceptable.
We used the /.autofsck mechanism to force a fsck after cold reboot, but
rc.sysinit still asks for human intervention. We then started to use the
/fsckoptions mechanism with the following content:
This makes fsck pass the ï¿½-yï¿½ parameter to fsck.ext2 to automatically respond
ï¿½yesï¿½, so we donï¿½t go into the interactive mode....
With the current rc.sysinit, fsck is called this way:
fsck -T -a -- -y /
which fails. If the parameter order is changed as showed above, the call will be OK:
fsck -T -a / -- -y
Using the /fsckoptions mechanism with this parameter order, ext3 never goes
into interactive fsck, even if we turn the machine off many times in the middle
of a fsck.
This is a problem only for the / fs, because the command for the other fss is
different and OK.
Added in CVS, will be in 7.20-1. Thanks!
RHL9 still doesn't have this fix.
Any plan to include it in RHL9 ?
Not at this point.