Wilbo
11-08-2002, 07:44 PM
My server seems to go down every day. The only thing I can seem to find in the logs right before it went down last is this error:
Nov 8 09:31:55 ip40-10 kernel: EXT3-fs error (device ide0(3,6)): ext3_free_inode: bit already cleared for inode 245767
Nov 8 09:31:55 ip40-10 kernel: EXT3-fs error (device ide0(3,6)): ext3_free_inode: bit already cleared for inode 245765
I'm running Red Hat 7.3 with the 2.4.18-3 kernel.
Anybody got any idea what this error is and how it can be remedied?
sitekeeper
11-08-2002, 08:51 PM
The error is just ext3's normal reaction to a fatal error detected in the filesystem, so that in itself isn't a worry. The cause of the problem it spotted is the worry...
I would suggest you run RUN fsck MANUALLY without -a or -p options. If your are not sure what you are doing let someone where your server is located do it... This could lead to loss of data!
Alan - Vox
11-08-2002, 09:58 PM
What are the errors in the log before the time of the crash? Try keeping top running until the server crashes and post the output here.
Wilbo
11-08-2002, 10:58 PM
Originally posted by SplashHost.com
What are the errors in the log before the time of the crash? Try keeping top running until the server crashes and post the output here.
There are no errors in the logs. It's crashed like 5 times and there are never any errors. It just freezes, then it's rebooted. The logs always show that it's humming along then a restart after the tech at the colo facility reboots it.
fmadison
11-13-2002, 01:50 PM
Hi,
When running a web server always always... use ECC memory.
Also you could have power supply problems as well.
And if you have a highly loaded server you may want to increase the amount of maxclients from 256 to 512. You may need to recompile Apache.
Last but not least your 4.00 am crash usually caused by slocate.cron in the /etc/cron.daily/ folder move it to another location.
Hope this helps.
-Frank
