[Gluster-users] How to rotate logs on glusterfs 3.1.0

[Phil Packer]

Hi, thanks, Craig...

I looked at that, and I realise that this will let me rotate the logs under /var/log/glusterd/bricks and that I can call that from logrotate if I want, but how should I rotate the logs from the glusterfs clients such as nfs.log?

I presume something like this would work (I've not tested it) and if so, it might be a good idea to drop a script into the RPM to avoid other people struggling with it?

---snip--
# Rotate client logs
/var/log/glusterfs/*.log {
sharedscripts
missingok
postrotate
/usr/bin/killall -HUP glusterfs > /dev/null 2>&1 || true
endscript
}

# Rotate server logs
/var/log/glusterfs/bricks/*.log {
sharedscripts
missingok
prerotate
for i in `gluster volume info all | grep Volume | cut -f3 -d" "`; do
/usr/bin/gluster  volume rotate $i
done
endscript
}
---snip---



Thanks

Phil

ps I don't use nfs in this set up, how can I stop the glusterfs daemon providing nfs service?

Thanks again

P


On 3 Nov 2010, at 10:22, Craig Carl wrote:

> Phil - 
>    Please check this page - http://www.gluster.com/community/documentation/index.php/Gluster_3.1:_Rotating_Logs
> 
> 
> 
> Thanks,
> Craig
> 
> -->
> Craig Carl
> Gluster, Inc. 
> Cell - (408) 829-9953 (California, USA)
> Gtalk - craig.carl at gmail.com
> 
> 
> From: "Phil Packer" <philp at layer3.co.uk>
> To: "Gluster General Discussion List" <gluster-users at gluster.org>
> Sent: Wednesday, November 3, 2010 2:22:23 AM
> Subject: [Gluster-users] How to rotate logs on glusterfs 3.1.0
> 
> Hi, folks,
> 
> Now the autofs issues have been sorted out (for which, thank you very much) I've finally upgraded my client's system to 3.1.0 but I now have some issues with logfiles.
> 
> 
> Currently running on Centos 5 and using a script in /etc/logrotate.d which used to work with gluster 2.n which no longer seems to work...
> 
> The logfile rotate script I've been using looks like this:
> 
> [l3admin at oy-imageserver2 glusterfs]$ cat /etc/logrotate.d/glusterfs 
> /var/log/glusterfs/*.log /var/log/glusterfs/bricks/*.log {
> sharedscripts
> missingok
> postrotate
> /usr/bin/killall -HUP glusterfsd || true
> endscript
> }
> 
> And currently the log files we have are:
> 
> [l3admin at oy-imageserver2 ~]$ cd /var/log/glusterfs/
> [l3admin at oy-imageserver2 glusterfs]$ ls -lt | head
> total 704144
> -rw-r--r-- 1 root root 208508659 Nov  3 09:08 nfs.log.1
> -rw-r--r-- 1 root root      8941 Nov  3 08:29 mnt-auto-shared.log
> -rw-r--r-- 1 root root   2927576 Nov  3 05:09 mnt-auto-shared.log.1
> -rw-r--r-- 1 root root         0 Nov  3 04:02 etc-glusterfs-glusterd.vol.log
> -rw-r--r-- 1 root root         0 Nov  3 04:02 nfs.log
> -rw-r--r-- 1 root root     21715 Nov  2 08:20 etc-glusterfs-glusterd.vol.log.7
> -rw-r--r-- 1 root root         0 Nov  2 04:02 etc-glusterfs-glusterd.vol.log.1
> -rw-r--r-- 1 root root    629632 Nov  2 04:02 mnt-auto-shared.log.2
> -rw-r--r-- 1 root root 146007238 Nov  2 04:02 nfs.log.2
> 
> 
> And as you can see the nfs log has been rotated but one of the processes is still writing to the old file...
> 
> Also in the bricks subdirectory, the log file isn't getting rotated at all.
> 
> Does anyone have a working logrotate script for RHEL/Centos??
> 
> Regards to all
> 
> 
> Phil
> 
> 
> -- 
> Director, Layer3 Systems Ltd
> Layer3 Systems Limited is registered in England.  Company no 3130393
> 43 Pendle Road, Streatham, London, SW16 6RT
> tel: 020 8769 4484
> web: www.layer3.co.uk
> 
> 
> 
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://gluster.org/cgi-bin/mailman/listinfo/gluster-users

-- 
Director, Layer3 Systems Ltd
Layer3 Systems Limited is registered in England.  Company no 3130393
43 Pendle Road, Streatham, London, SW16 6RT
tel: 020 8769 4484
web: www.layer3.co.uk