Here’s what’s happening: In the last few weeks, any database that is set to “Immediate” for indexing is not indexing AT ALL. We’ve had to set up Program Docs to run UPDALL on those databases that are most important to get indexing to work.
Now, this is happening on ALL our servers. And ALL databases that have document input (mail in databases index fine). So, we know it’s those databases that have document creation by users (these are web-only databases).
We’ve added DEBUG params to some of the servers to watch the Indexers. Bottom line: Indexer is NOT kicking off for those databases that have unindexed documents.
Any ideas? We’ve increased the amount of updaters to three, which did no good. Deleted and recreated indicies, which also did no good.
Immediate indexing does not index immediately because the server waits for the right moment to do it. The problem is that the right moment barely seems to exist due to the algorithm used.
We have experienced the same situation and the only solution we’ve found is to set the update frequency to “Hourly”, otherwise the databases never got indexed.
Thanks for the info. IBM also suggested we go to an hourly update schedule, but I’m not satisfied with that. If IBM/Lotus can’t get their “immediate” indexing to work, especially on large databases, then get rid of it.
I don’t mean to be dense, but you are talking about your full text indices, right? Have you turned on any indexer logging or debugging to see whether the indexer task itself is running reasonably? For this, I might scale back to 1 indexer just to make troubleshooting easier.
Yes, I’m talking about full text indices. We have turned on several debugging parameters. What’s funny is we see the updater task starting, running and completing normally.
However, it skips some database, usually LARGE (50MB or more) databases. As I said before, we’ve deleted the indices for these databases and recreated them to no avail.