Maintain view indexes on cluster server

We run our mail servers in a active-passive cluster setup where the mail files on the backup cluster server are never accessed by a user until the primary node goes down. The problem is when the backup server is hit with all the users at once, it slows to a crawl trying to build all the view indexes in the mail files.

We have run updall -c to build the indexes on the backup server and set the Default_Index_Lifetime_Days INI variable to maintain the indexes for a certain number of days. At some point though the server will purge the view indexes since a user hasn’t accessed it. Does anyone know of a way to maintain the on-disk view indexes via an agent or other server process?

I was told by IBM that updall -V wouldn’t reset the lifetime counter and it looks like view.Refresh in LS won’t work either.

Thanks,

-Kevin

Subject: Maintain view indexes on cluster server

Kevin,

I’m just giving an educated guess here, not a voice of experience, but I think I would have a program doc that runs nightly which would update a selected set of views, not all views, so that I compromise between performance and wasted disk space.

For the nab, I’d do the name and group lookup views, for example.

regards,

raphael