Last friday my users started receiving warnings on their mail clients (and webmail) that their mail could not be sent due to lack of disk space. Alarming, curious, but it happens. Funny thing was i has just gone thru all my boxes to make sure nothing was short on space probably a week before. there was some stuff i could move onto SAN to make room, and I created about 11G. On Monday the 11G was gone. I made a few G more. Today it is gone again.
Zimbra is the only application running on the physical disc on this machine. everything else is ln -s linked to the SAN. I cannot for the life of me figure out where the space is going. It is not a constant ticking away of space but an ebb and flow for awhile. no change for awhile and then a giant bite. then ebb and flow, stable, giant bite.
My questions are these:
1) any idea what in zimbra could be doing this? Large attachments or messages? some sort of logging feedback loop?
2) is it a definite no no to move /opt to my SAN and ln -s link? At least this way i have enough space to keep mail flowing while i figure out what is happening. I've read the scuttlebutt about updates not working across the links and im not too worried about this. We are going to build a commercial version of ZCS on a new server in the near future and migrate to it. No updates for our version.
3) Also, my logger files are pretty huge, and i could use that space. for the time being. I read that this code works
But i want to be clear.su - zimbra
stop logger db (mysql processes)
rm -rf /opt/zimbra/logger/db/data
Is not really code. what i want to do is stop mysql, correct?stop logger db (mysql processes)
Delete the directory (or in my case move it to the SAN so i can look it over)rm -rf /opt/zimbra/logger/db/data
run this script which re-creates the dir structure for logging./opt/zimbra/libexec/zmloggerinit
Start mysql, start logger.
is this the gist?
I've already leaked 600MB since this morning, and have about 900MB left on the disk, so please help and thanks in advance. Time is of the essence.