On 03/12/2011 12:01 AM, Greg Freemyer wrote:
On Fri, Mar 11, 2011 at 4:52 PM, Vahis
wrote: On 03/11/2011 11:39 PM, jdd wrote:
Le 11/03/2011 20:17, Vahis a écrit :
On 03/11/2011 08:58 PM, jdd wrote:
jdd nb: not to say every 20 starts, full fsk...
That can easily be changed:
not sure it's a good idea on a server. On this box, I *need* a safe reboot, better wait than have a corrupted filesystem... also there is an UPS question regarding boot/stop
But if the full check is automatic after some number of mounts you don't have the control over it.
I prefer the check when _I_ want it, not the system.
The journal is checked at every boot and if there's a problem(s) you'll be prompted to do a manual check, which is fine.
I'm talking about automatic _full_ check which I dislike and disable.
Ted Tso (the ext4 maintainer) recommends for servers:
1) No auto fsck
2) Put FS on LVM
3) Routinely (from cron) make a snapshot
4) fsck the snapshot and check for errors
5) If errors, notify admin so they can schedule down time to run fsck and fix them.
Greg
Now there you might have a point for me to start using LVM :) I've never seen its benefits, but snapshots could be it, thanks :) Vahis -- http://waxborg.servepics.com openSUSE 11.2 (x86_64) 2.6.31.14-0.6-default as main host, with guests: openSUSE 11.4 (x86_64) 2.6.37.3-16-desktop "Tumbleweed" openSUSE 11.4 (x86_64) 2.6.37.1-1.2-desktop -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org