theone Posted February 5, 2014 Share Posted February 5, 2014 I am getting the following email from my unraid server every minute: Subject: cron for user root /usr/lib/sa/sa1 2 1 1> /dev/null Cannot write data to system activity file: No space left on device [code] I ran "df" and got: [code] root@Tower:~# df Filesystem 1K-blocks Used Available Use% Mounted on tmpfs 131072 131072 0 100% /var/log /dev/sda1 7819328 575232 7244096 8% /boot /dev/md1 1953454928 1878535224 74919704 97% /mnt/disk1 /dev/md2 976732736 959285160 17447576 99% /mnt/disk2 /dev/md4 1953454928 1914671804 38783124 99% /mnt/disk4 /dev/md5 2930177100 1494300852 1435876248 51% /mnt/disk5 /dev/sdd1 244191092 128747676 115443416 53% /mnt/cache shfs 7813819692 6246793040 1567026652 80% /mnt/user It seems that my tmpfs is 100% FULL !!! This is my syslog file: 348 -rw-r--r-- 1 root root 356352 2014-02-05 17:46 syslog Why is this happening? This is not the first time. It usually happens ~40 days after server is up. currently ~40 days. edit: After deleting the syslog: root@Tower:~# df Filesystem 1K-blocks Used Available Use% Mounted on tmpfs 131072 130752 320 100% /var/log /dev/sda1 7819328 575232 7244096 8% /boot /dev/md1 1953454928 1878535224 74919704 97% /mnt/disk1 /dev/md2 976732736 959285160 17447576 99% /mnt/disk2 /dev/md4 1953454928 1914671804 38783124 99% /mnt/disk4 /dev/md5 2930177100 1494300852 1435876248 51% /mnt/disk5 /dev/sdd1 244191092 128747696 115443396 53% /mnt/cache shfs 7813819692 6246793040 1567026652 80% /mnt/user How can I see what is taking all the space in tmpfs? Quote Link to comment
hoek Posted February 5, 2014 Share Posted February 5, 2014 It's the System Stats plugin for SF and Dynamix. I've never had issues with /var/log filling up until enabling that plugin. I imagine you can direct logging for sa to go somewhere else or set it to roll over but I just haven't looked into it. ls -l /var/log/sa hmmm just checked mine again and it looks like it does roll over after sa31. Still might be worth looking into moving the logging to a different place... Quote Link to comment
theone Posted February 5, 2014 Author Share Posted February 5, 2014 I ran: root@Tower:~# du -a /var | sort -n -r | head -n 10 143116 /var 130756 /var/log 128952 /var/log/sa 11912 /var/cache 11892 /var/cache/samba 11360 /var/cache/samba/locking.tdb 6312 /var/log/sa/sa31 6312 /var/log/sa/sa30 6312 /var/log/sa/sa29 6312 /var/log/sa/sa28 and indeed it is SA. It seems that everyone should have ha this issue, I haven't come across it on the forums before. Quote Link to comment
hoek Posted February 5, 2014 Share Posted February 5, 2014 I think given a certain combination of plugins and activity for your particular setup it may affect you. Quote Link to comment
dirtysanchez Posted May 14, 2015 Share Posted May 14, 2015 I am having this same issue since upgrading to v6. It is indeed SA filling tempfs and then logging stops. I have been manually deleting the saXX files once tempfs fills up so that logging can resume. Wondering if this is on LimeTech's radar as an issue? Quote Link to comment
dmacias Posted May 14, 2015 Share Posted May 14, 2015 Try this. It will increase your tmpfs. I have it in my go file but you can just execute it and put it in the go file for reboot. There a thread about it somewhere. # resize tmpfs mount -o remount,size=256m /var/log Quote Link to comment
trurl Posted May 14, 2015 Share Posted May 14, 2015 Try this. It will increase your tmpfs. I have it in my go file but you can just execute it and put it in the go file for reboot. There a thread about it somewhere. # resize tmpfs mount -o remount,size=256m /var/log Been in my go file since before the last upgrade. Quote Link to comment
dirtysanchez Posted May 15, 2015 Share Posted May 15, 2015 Good to know, I had missed that. I'll do it now, thanks. I assume 256MB for tempfs provides enough space for the SA logs to rotate through sa31 without filling up tempfs, and therefore resolves the issue? Quote Link to comment
dmacias Posted May 15, 2015 Share Posted May 15, 2015 Good to know, I had missed that. I'll do it now, thanks. I assume 256MB for tempfs provides enough space for the SA logs to rotate through sa31 without filling up tempfs, and therefore resolves the issue? For me it did. I think mine peaked at 140-150. That was on v6 and last beta had a libvirt debug flag set or something that cause the libvirt log to get too big. If not just raise it more. Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.