March 13, 201214 yr Hi guys, Require some serious help here. I have a server, running 5.0 Beta14, its been running perfectly and recently I started upgrading drives from 1TB's WD's to 2TB WD's, two at a time, obviously installing one then rebuilding and thereafter installing the next. However I have now got a problem whereby after the second set of drive upgrades my server would run and then crash after about 24 to 30 hours (this seems to remain constant), the web interface is not available and the server completely drops off the network. Reboot the unit and everything seems fine for a while after. I ran reiserfsck and found one of the drives to be corrupted and then ran reiserfsck --rebuild-tree. This seemed to sort the problem until it came up again. I have now changed out the motherboard and memory and the server ran for two days with not a hassle until it returned to this same problem. I have not experienced any issues with both of my servers for about 2 years so I am a serious newbie at going any deeper than the web interface. In another post someone mentioned checking the flash drive for problems, could this possibly cause this? I am baffled by the time period it takes to crash. I won't be able to test this theory until the morning. I appreciate any ideas, thoughts, suggestions etc. Can I still pull the syslog from the crash after rebooting the machine? Thanks again!
March 14, 201214 yr Run men test overnight. See here: http://lime-technology.com/forum/index.php?topic=9880.0
March 14, 201214 yr Author Ran mem test with no problems. Syslog is attached but this is one done after rebooting from a crash and starting the array. As mentioned before I am unable to access the syslog once the server has crashed. The server has been running now for an hour and a half with no problems, and I have run SMART tests on all the Drives with no errors. Again thanks for any help offered. syslog.txt
March 14, 201214 yr If the server drops from the network hook up a keyboard and monitor(a pain I know) to the server and do this: cp /var/log/syslog /boot/syslog-MM-DD-YYYY.txt Obviously replacing the MM-DD-YYYY with the current date. This way you can keep them seperate. I had the same problem for a while, not having it anymore. I disabled cache_dirs and its seemed to sort itself. I don't think my problem is in any way related to yours, just saying what I did. Every time my server has crashed/dropped from the network it was still responsive at the server itself so I'd give that a try, atleast to get a syslog that will show whats happening before and during the crash.
March 15, 201214 yr Run men test overnight. See here: http://lime-technology.com/forum/index.php?topic=9880.0 Use the telnet/tail method described in the link to capture the syslog.
Archived
This topic is now archived and is closed to further replies.