Recommended Server Maintenance

HCL Domino Domino Forum
1

Recently, a considerable number of employees have encountered corrupt mail databases. Running a “compact -c” usually fixes the problem. Can anyone recommend or suggest…

  1. What maintenance (e.g. specific commands) should we be running as a preventative measure?

  2. How frequently we should be performing maintenance?

  3. What’s the max recommended mail database size (or threshold) if we start to enforce limits (e.g. keep the mail file under 500Mb)?

Server background info… We’re running two domino servers with v6.5.2. Both are Sun servers with 7,500 mail files per box. Some of the mail databases are rather large, 1Gb or larger (sigh).

Thanks in advance,

Mark

2

Subject: Recommended Server Maintenance

Well, depending on your company’s standpoint on mail storage and retention, the first thing I would do would be to implement a mail file quota. We use 200Mb, and then have a couple of other levels depending on needs of the user.If you’ve got 7500 mail db’s per server it’s gonna take a long time to compact them all if they’re large databases. I have approximately 2200 db’s per primary server and I run 3 compact program docs simultaneously and still run into some instances where compact is still running the next morning, 8 hours later. Granted, I’m not running the “-c” option, and I’m in a W2k environment on blade servers but that should give you a little bit of an idea.

You might also want to consider running fixup with the “-J -O” options as well.

3

Subject: RE: Recommended Server Maintenance

Thank you for the prompt reply. This helps.

I’m a bit knew to server admin. Most of my background is with application development. Could you possibly provide some additional information pertaing to the “3 compact program docs” (what options to set)?

Also, this may be a silly question… should I compact the databases with the server up and down?

Thanks Again, Mark

4

Subject: RE: Recommended Server Maintenance

I have 3 primary mail directories on each server and then 3 secondary (cluster failover), for a 6 server cluster. I have transaction logging enabled on each server as well.As I’m really only concerned with backing up the primary mail files on each server, I run compact with either the -B or -b options depending on what day of the week it is for the server’s full backup or incremental backup of mailfiles and transaction logs.

I run all of my compacts with the servers up and running, I can’t have servers down for any extended length of time since I have a lot of DWA and webmail users and do not have ICM implemented yet.

As far as the options you’ll want to set, that depends greatly on your own environment. If you have multiple mail directories as I do, you can create program docs for each of them and run them concurrently, or you could create one and just let it run through the entire server. One doc will take longer but it will use less CPU than multiple docs, it depends on where your preferences lay in that regards.

The admin guide has listings of all of the various switches/options available for compact,

the ones I find to be most useful are generally the -c (copy style, helps resolve corruption) and the -S # (where the # is the percentage of white space the db must be above for compact to occur).

Hope this helps.

5

Subject: RE: Recommended Server Maintenance

I’ve found that running a compact with the -S5 switch greatly reduces the time that compact takes on a server with 400GB of mail files.

We had problems with compact starting during the weekend, but not finishing on Monday morning until we started using this switch.