[CentOS] unattended fsck on reboot

Wed Jan 6 19:17:18 UTC 2010
Robert Heller <heller at deepsoft.com>

At Wed, 06 Jan 2010 11:45:46 -0600 CentOS mailing list <centos at centos.org> wrote:

> 
> On 1/6/2010 11:19 AM, Alan McKay wrote:
> > Hey folks,
> >
> > I searched the list archives and found this :
> >
> > echo "AUTOFSCK_TIMEOUT=5">  /etc/sysconfig/autofsck
> > echo "AUTOFSCK_DEF_CHECK=yes">>  /etc/sysconfig/autofsck
> >
> > http://lists.centos.org/pipermail/centos/2006-November/029837.html
> > http://lists.centos.org/pipermail/centos/2009-September/thread.html#81934
> >
> > Will this do all disks?
> >
> > I want to do a reboot of a couple of systems during our maintenance
> > window and fsck them, but would rather try it from home first and not
> > go to the data center.  Then of course rush there like a madman if
> > they don't come back up :-)
> >
> > There was a suggestion in the 2nd thread above that with ext3 this
> > should not be required with proper hardware (my paraphrase).   I'm
> > using all IBM stuff - x3550, x3650, x3800 and some of the earlier
> > models like x330.  I can't imagine this being an issue.   But
> > nonetheless I do have some issues on a couple of systems that look
> > like they need fsck'ing
> 
> It will happen by itself at some default interval.  I've forgotten 
> exactly what the timing is but it is infrequent enough that it always 
> takes me by surprise when it takes an extra 10 minutes for a remote 
> system to come back up.

There are two metrics used:  number of times a FS is mounted and number
of days since last fsck/mount.  For machines that don't get rebooted
often (eg servers) the 'number of times a FS is mounted' almost never
kicks in and the 'number of days since last fsck/mount' does.


-- 
Robert Heller             -- 978-544-6933
Deepwoods Software        -- Download the Model Railroad System
http://www.deepsoft.com/  -- Binaries for Linux and MS-Windows
heller at deepsoft.com       -- http://www.deepsoft.com/ModelRailroadSystem/