system "freeze" while periodic daily running

P

PeWiWa

I'm running: FreeBSD 9.0-RELEASE

Hi,

this morning I noticed my system was not running any more. I was able to login, but each command took about a minute or more. A ps ax showed a running periodic daily within its security-scripts that started at 03:00. There were 2 or 3 find's searching for set-uid-files - but it was 09:30! I had to hard-reboot (means: switch off) the system, after that it ran perfectly again.

The logs give no idea of what was happening. As always cron logged the start of periodic daily in /var/log/cron, but nothing else. Nothing in messages. The days before the system was running without problems.

The only running app in this system is a jailed apache22 (normally doing nothing :)).

All in all: I have no hints and no idea. Does anybody else have?

Peter
 
Those scripts scan most file systems that are mounted, so if its bogging your system down it might indicate a faulty storage device. Any kernel disk errors when this happens? Are you performing regular SMART self-tests on your disks?
 
@aragon: no, there are no errors in the logs

The filesystems are:

Code: 
server# mount
/dev/ada0s1a on / (ufs, local, journaled soft-updates)
devfs on /dev (devfs, local, multilabel)
/usr/local/www on /usr/jails/www/usr/local/www (nullfs, local, read-only)
devfs on /usr/jails/www/dev (devfs, local, multilabel)

BTW: I don't know if it is of interest. I mounted a part of the host's www-area into the jail as readonly (for security reasons).

Peter
 
I'm facing similar problem. The periodic jobs runs and takes all the memory because it scans a large filesystem. Finally the find process gets killed and takes all my sessions including the nfs server.

Is there a tuning guide for periodic.conf?

Thanks.
 
I up this thread instead of opening mine.

My periodic run takes about 12k read IOPS for around 6 hours. I want to disable periodic on some directories. How can I do that?
 
You must log in or register to reply here.
Back
Top