Admin Backup/Transfer freeze my server

[perrera]

I dont know why, but every time a scheduled an Admin Backup run, my server freeze and have to reboot manually.

Tried 3 diferents disks, thinking it was a hard disk problem.

Anyone with a clue?
Thanks.

 

[zEitEr]

You might need to check your hardware (don't know what you mean by "Tried 3 diferents disks", if you did not change temp directory in directadmin conf it might have no big sense), and of course read system logs. You gave so little information, that's hard to guess anything more.

 

[perrera]

Hi thanks for the answer,
I am getting this in my daily cron now, strange thing.
Is anything related?

/etc/cron.daily/tmpwatch:

error: failed to lstat /tmp/sess_5bc8dcd44ced116a2375f741b0b8331d: Input/output error
error: failed to lstat /tmp/sess_5bc8dcd44ced116a2375f741b0b8331d: Input/output error

 

[mr.applesauce]

Is /tmp full?

What is the output of:

df -h

 

[perrera]

hi,
here it is:

[root@ideaweb3 ~]# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/VolGroup00-LogVol00 435G  106G  307G  26% /
/dev/sda1              99M   33M   62M  35% /boot
tmpfs                 7.6G     0  7.6G   0% /dev/shm
/var/tmpMnt           962M   20M  893M   3% /tmp

 

[perrera]

Any posibility that de TMP partition is to small?
I have clients over with near 10GB of disk usage.

 

[mr.applesauce]

The size is fine and its barely full. Session files shouldnt be using much space anyways. Might want to check google for the error to see if you can find any answer. I am not sure what the problem would be. Considering you tried 3 different hard drives, the problem might be with your disk controller of your motherboard.

 

[toml]

The lstat() system call requires read/execute permissions on all paths leading up to the file being checked. In this case if /tmp doesn't at least have -r-x------ then that could be the issue. What does `ls -ld /tmp` show?

 

[perrera]

Thanks for that reply, here is the output:

[root@ideaweb3 ~]# ls -ld /tmp
drwxrwxrwt 13 root root 20480 Aug  3 15:39 /tmp

I manaullay delete the file of the error: /tmp/sess_

The freeze seems to happend when I backup all users at the same time (120 users aprox.). Partially backups is working fine.

 

[toml]

That looks fine, there has to be something else causing the freeze.

The lstat problem could possibly be ignored, if the filesystem checks out. tmpwatch was checking it to see if it can be deleted after a certain period of time.

Have you tried monitoring the system while doing those backups? Like is there enough free memory or did the system start to swap. Some systems crash when you run out of free memory or they run so slow they appear to freeze when they start thrashing (aka swapping).