Hi all,
I've had a nagging problem for ages now, which I think I've just worked out. Firstly, can someone tell me if I'm talking absolute rubbish, and secondly, if I'm not, can someone help me find a resolution.
Here's the scenario.
We have 2 DB2 systems, a test and a production. The test goes over the LAN, and the Production goes LAN Free.
The test system goes straight to the diskpool, which then gets migrated to tape at 05:00 every morning. Obviously being LAN Free, production goes straight to tape.
The main issue we've been having, is the incremental backups which run every night and every lunchtime. The production box picks up a tape, and happily goes on its way. The test system writes about 8 gig to disk (which is to be expected). Here's where the fun starts. In the last 5 minutes or so of its backup, the test system decides that it needs a tape, and pre-empts the mount point that the production server is using. This cancels the session for the production server, and the backup fails.
My thinking is - the bulk of the backup for the test system goes to disk, but when it gets to the "Include logs" part of the backup, its trying to put those straight to tape.
I think I have gotten round this by increasing the number of mount points that the test system can use (I will have to wait until tomorrow to check if this actually makes a difference.), so in theory it will mount a second tape if the production box is using the first. However, if this does work, I'd still like to get it writing solely to disk, otherwise I might as well just write the whole lot straight to tape anyway.
Any thoughts?
I've had a nagging problem for ages now, which I think I've just worked out. Firstly, can someone tell me if I'm talking absolute rubbish, and secondly, if I'm not, can someone help me find a resolution.
Here's the scenario.
We have 2 DB2 systems, a test and a production. The test goes over the LAN, and the Production goes LAN Free.
The test system goes straight to the diskpool, which then gets migrated to tape at 05:00 every morning. Obviously being LAN Free, production goes straight to tape.
The main issue we've been having, is the incremental backups which run every night and every lunchtime. The production box picks up a tape, and happily goes on its way. The test system writes about 8 gig to disk (which is to be expected). Here's where the fun starts. In the last 5 minutes or so of its backup, the test system decides that it needs a tape, and pre-empts the mount point that the production server is using. This cancels the session for the production server, and the backup fails.
My thinking is - the bulk of the backup for the test system goes to disk, but when it gets to the "Include logs" part of the backup, its trying to put those straight to tape.
I think I have gotten round this by increasing the number of mount points that the test system can use (I will have to wait until tomorrow to check if this actually makes a difference.), so in theory it will mount a second tape if the production box is using the first. However, if this does work, I'd still like to get it writing solely to disk, otherwise I might as well just write the whole lot straight to tape anyway.
Any thoughts?