This is a request as a hardware support person for a client using a database created for them (not by me). Originally the database was created in AREV, then updated to work in OI. We have had a succession of problems with more than one user. Now with OI there seems to be a similar problem. The original database was used on a PC as a server/workstation. No problem as long as only one user was on. With two users there were problems mostly reporting index errors or lockups. It ws assumed that the problem was the workstation so the database was transferred to another PC. Same deal. Eventually we introduced a NT4 Server with Service release 2 installed. The network is 10 Base with two Win 95 workstations.

The OI is ver 3.6, Network driver is 1.5.

With one user on there is generally not many problems but sometimes it reports that there are errors with indexes, which it requires to be rebuilt. The database manager doesnot seem able to fix these and reports errors. The clean up must be done in System monitor!

With one user on and a second user coming on line if searches are carried out then the whole lot locks. If one of the users is forced out then the other may carry on. (The user forced out cannot restart OI as it returns a failure to start message and the PC must be rebooted !)

While the database is obviously good I suspect a few underlying problems and would appreciate some help.

As a footnote, in the AREV release there was the ability to automate with a macro the ability to check the data using LHverify, but this is not apparently available in the Windows system. Any thoughts.

Bill,

If everything is setup properly the problems you are describing should not be happening. BTW, where is the original programmer for this database? If this person is no longer available I would strongly suggest contacting someone who can provide direct support, especially if the "worst" case scenario exists.

Since this problem has apparently existed pre-OI (i.e. AREV) I would suspect the network configuration. In any event, here are a couple of fundamental issues I'd check out first:

1. Make sure basic locking is taking place. Run the application on two machines and try to pull up the same record on both machines at the same time. One machine should get the record and the other should give a warning that the record is in use but that it can be viewed in "Read Only" mode. If this doesn't happen, then you've definitely got a problem and this would be the cause for all the indexing corruptions.

2. Make sure that the indexes are rebuilt from scratch according to Technical bulletin #56.

3. Check all the tables, including system tables, for GFE's using the Database Manager.

4. Run the NETDRV.EXE program stored in the directory that OI resides in. Make sure that it indicates a network driver in the section "Current Network:" and make sure the "Maximum user:" count is greater than 1.

Pursue these things and let us know what you come up with and we'll see if we can't nail this one down for you.

dbakke@srpcs.com

SRP Computer Solutions

Bill,

It sounds like you have either locking issues or a configuration issue that could cause GFE's. Please verify that locking works. Are any of the problems that you have encountered GFE's? If so, it could be caused by a variety of things including reboots of the workstations or network errors.

Cameron Purdy

Revelation Software

Thanks for your message. I have tried files locks between the two users and that seems to be OK.

We are using NETBUI as the standard protocols. Is there any preference by Open Insight ?

We do get GFE's but these only show up when we do checks, they do not report as such when operating the programme. You may have seen a message that Mat Crozier put up on the Network section, which also relates to our problem.

Cheers

Bill

Don

Thanks for the help.

We have checked the record locks and they seem to be working OK!

The individual programmer is still available, and he has put up a note on the forum under the network section (Mat Crozier)

We have checked all the things that he and I can think of re the network or the programme. The next option seems to be that installation of the Performance Pack which from resading another message seems to reduce the throughput. As we have a NT server and Win95 workstations do we need the NPP which is universal or NTPP?

Regards Bill

Hi Bill,

You're already using the NPP 1.5.0.0 (aka All Networks Driver). By NTPP you mean the NT Service? I believe we did check the frame type but better be sure as per Cameron's responce.

Also, could you please go into Database Manager and select Database, Enironment Settings, Indexes, and tell me what Seconds Before Update is?

Matt.

Matt was right - changing indexing from default 5 sec to 10 minutes and suddenly no sluggish computers and (consequently) no GFE's due probably because people crashed computer because they had thought it had hung when it was just indexing!

I hear that these users are now VERY HAPPY.

View this thread on the forum...