I have a number of servers at different clients that are experiencing this problem. During the periodic full-text updates, the disk is hammered and the functionality of the server is seriously affected. I have been forced to change the update interval to ‘Daily’, however this leaves inaccurate indexes until the next day.
I am aware of the ini variable ‘FTUpdate_Idle_Time’, but that only works between databases. This happens on single database that can take 3-4 minutes to update less than 100 changes to normal sized documents. Turning attachment indexing off does not help either. Also, I am aware there is a fix in 7.0.2 for large indexes, but that has no noticeable effect either.
I am fine with the lengthy process, I just cannot have the server be taken to 100% in the middle of the day. Is there any way to throttle the update process?
Yes, I have tried deleting and rebuilding the indexes.
This happens on a number of databases on different servers with different attributes. An example is a database with 200,000 documents @ 5Gig, certainly within reasonable size limits.
Also, I am certain that it is the full-text Update process. Domino reporting is very specific, you know how to pull it. I have been trying to correct this issue for a couple years. I just thought, I’d ask the question to see if anyone has a solution. If not, no problem.
I have been monitoring one of the servers and have identified that swapping memory (paging) is the root cause of the drive usage. This is strange as the server has 4 gig of ram, over the amount that Domino can recognize on Windows.
Is there anyway to prevent nupdate.exe from building large amounts of cache that needs to be swapped in when it periodically runs? I would rather the full-text update process take 3 times as long than tie up the server as it does.
My option seems to be limited to moving the page file to a separate drive. Unfortunately, I do not think I have the room in all of the servers that are affected by this issue.