We have clients running AREV 3.12 applications with several hundred data files at times. These systems are running on Novell 3.12, 4.1plus, and Window NT Server. Some of the Novell sites use the NLM and all of the NT Server site use at least the NPP and most use NT Serve. Indexing is used to improve perform and filter lookup data for popups because of popups size limitation.

The problem is with background indexing. Most clients have between 5 - 40 active user that leave AREV running all day. When several of the workstations go into background indexing there seems to be a lock conflict that at least appear to the user make there workstation is locked. The workstation actually updating indexes will stop as soon as a key is pressed, but the other workstations will not response to the keyboard and stop indexing as you would expect. I assume this is because the workstation has not timed out waiting on locking on the server. It appears to remain that way until the workstation performing index updates completes it's work or the user presses a key.

The users normal response to this apparent locked workstations is to reboot the DOS session. Normally we do not get corrupt indexes when they do this, so I'm assuming the workstation was not actually updating indexes. If the user gives the workstation enough time control will normally return to the users, but most are too impatient to wait.

Is there a way to time out the wait on network locks to release the workstation sooner. The user needs to get feedback from AREV sooner.

I'd set up a dedicated indexer. That would solve the problem. OK, mask the problem, but it would take care of it.

[email protected]

Sprezzatura, Inc.

As Aaron said, set up one or more dedicated indexers. Why? Because that's what works best.

View this thread on the forum...