We have a user uing the NPP with OI on an NT network. They have one workstation at another site via a WAN. Despite much cajoling they refuse to go NT Service.

They report intermittent retained locks at various workstations, these locks being held until the user logs out of OI.

Another separate user reported this behaviour also so before we trawl for orphaned lock statements in code (we effectively inherited this system many moons since) could anyone else comment whether they have seen this behaviour?

Thanks!

The Sprezzatura Group

World Leaders in all things RevSoft

We've had a site that have just upgraded to the 2.1 NT Service (on 23/9/02) and are now getting similiar behavior. It is a 10 user site that was not getting locking problems before the 2.1 NT Service upgrade.

Our software has not changed, so it seems to be a problem in the service or NPP. Same behavior - locks are randomly held and are not released until the user logs out of OI. We have a second site using the 2.1 Service who have been having similiar problems.

Are you still getting the problems at your site? Has anyone else encountered this, or does anyone have any suggestions?

Thanks,

Paul Rowe

We've managed to narrow this problem down further. We have three sites using the 2.1 Service with Win 2000 servers. All immediately began getting locking errors when they upgraded to the 2.1 service.

One site has set the registry on the server to force the Service to only use NamedPipes. The locking errors have now gone from that site, so it seems to be a problem with using TCP/IP protocol.

Other info:

Win NT/2000/XP workstations all seem affected.

It just seems to be random locks that are not removed. It was reported to us around three times a day by the 10 user site, but the problem may have been happening much more often - they only notice if they try to read the same record again later and it is still locked.

Our application does a large amount of locking - each 'record' viewed involves around 2-5 locks, as the information presented on screen is read from several related linear hash records.

We also get a second problem with the 2.1 Service. User logins are not always cleared (possibly after an abnormal end). This results in an error on login

FS415 - Cannot select on SYSLOGINS

We've tried a suggestion from Don Bakke to check records in SYSLOGINS and delete any that are not locked, but this has not resolved the problem. Again, the problem may be specific to TCP/IP, but we haven't tested the NamedPipes site for enough days yet to be sure of that (as the error is less frequent).

Is there anyone else that has either of these errors with the 2.1 Service?

Thanks for the various suggestions so far, we always appreciate the help.

Paul

View this thread on the forum...