but
[2014-08-26 11:45:35.315974] I
[afr-self-heal-common.c:2868:afr_log_self_heal_completion_status]
0-HA-FAST-PVE1-150G-replicate-0: foreground data self heal
is successfully completed, data self heal from
HA-FAST-PVE1-150G-client-1 to sinks
HA-FAST-PVE1-150G-client-0, with 16108814336 bytes on
HA-FAST-PVE1-150G-client-0, 16108814336 bytes on
HA-FAST-PVE1-150G-client-1, data - Pending matrix: [ [ 0 0
] [ 348 0 ] ] on
<gfid:e3ede9c6-28d6-4755-841a-d8329e42ccc4>
something wrong during upgrade?
I've got two VM-s on different volumes: one with HD on and
other with HD off.
Both survived the outage and both seemed synced.
but today I've found kind of a bug with log rotation.
logs rotated both on server and client sides, but logs are
being written in *.log.1 file :)
/var/log/glusterfs/mnt-pve-HA-MED-PVE1-1T.log.1
/var/log/glusterfs/glustershd.log.1
such behavior came after upgrade.
logrotate.d conf files include the HUP for gluster pid-s.
client:
/var/log/glusterfs/*.log {
daily
rotate 7
delaycompress
compress
notifempty
missingok
postrotate
[ ! -f /var/run/glusterd.pid ] || kill
-HUP `cat /var/run/glusterd.pid`
endscript
}
but I'm not able to ls the pid on client side (should it be
there?) :(
and servers:
/var/log/glusterfs/*.log {
daily
rotate 7
delaycompress
compress
notifempty
missingok
postrotate
[ ! -f /var/run/glusterd.pid ] || kill
-HUP `cat /var/run/glusterd.pid`
endscript
}
/var/log/glusterfs/*/*.log {
daily
rotate 7
delaycompress
compress
notifempty
missingok
copytruncate
postrotate
[ ! -f /var/run/glusterd.pid ] || kill
-HUP `cat /var/run/glusterd.pid`
endscript
}
I do have /var/run/glusterd.pid on server side.
Should I change something? Logrotation seems to be broken.
