My configuration
================
I'm supporting a Networker 5.1 installation on two HP-UX 10.20 servers, one of which is set up as a Storage Node. Both have a Storagetek 9714 jukebox attached, the main server has 6xDLT700 drives, and the Storage Node has 5xDLT4000 drives.
The problem
===========
During the daytime, Networker stops responding and you get repeated "malloc failed, retrying..." messages in the daemon.log. Networker will shutdown and restart ok, and on restart it reports that the media db is bad and needs scavenging, or that /nsr/mm cannot be mounted because of a bad database header. The scavenge/rebuild is always successful, but it's obviously very annoying. Luckily it has only happened about 4 times during the daytime at the moment, never during our busy overnight backup schedule. I can't match up the problems with anything that is happening at the time, although we do a small amount of relabelling of tapes each day. We also restart Networker every weekday at 5pm, so the media db is getting checked each time Networker starts up.
I know this is a really old setup and needs to be upgraded, but unfortunately that's not an option for me right now. We have a new backup solution waiting to go (Veritas Netbackup) and I'm just trying to apply bandages to our Networker setup to keep it ticking over until we can move everything over to Netbackup.
Does anyone have any ideas what might be causing the media index to get corrupted, and what I can do to prevent it?
Many thanks in advance.
================
I'm supporting a Networker 5.1 installation on two HP-UX 10.20 servers, one of which is set up as a Storage Node. Both have a Storagetek 9714 jukebox attached, the main server has 6xDLT700 drives, and the Storage Node has 5xDLT4000 drives.
The problem
===========
During the daytime, Networker stops responding and you get repeated "malloc failed, retrying..." messages in the daemon.log. Networker will shutdown and restart ok, and on restart it reports that the media db is bad and needs scavenging, or that /nsr/mm cannot be mounted because of a bad database header. The scavenge/rebuild is always successful, but it's obviously very annoying. Luckily it has only happened about 4 times during the daytime at the moment, never during our busy overnight backup schedule. I can't match up the problems with anything that is happening at the time, although we do a small amount of relabelling of tapes each day. We also restart Networker every weekday at 5pm, so the media db is getting checked each time Networker starts up.
I know this is a really old setup and needs to be upgraded, but unfortunately that's not an option for me right now. We have a new backup solution waiting to go (Veritas Netbackup) and I'm just trying to apply bandages to our Networker setup to keep it ticking over until we can move everything over to Netbackup.
Does anyone have any ideas what might be causing the media index to get corrupted, and what I can do to prevent it?
Many thanks in advance.
