Network errors

[HotEye]

We are using Novell 6 and a Clipper 5.2e/Blinker 7 application.

Over the last week or so, we have been getting DOS ERROR 2 with a RETRY option, upon retry it opens the file and all is OK. Sometimes it takes more than one retry to work.

This is happening on some DOS systems and some Windows 98 systems.

I think it may be to do with network timeouts, but how are these changed? It seems odd that it is a receint error, the server has not been changed in about 4 or 5 months.

The server utilisation is usually only around 1 to 10%, although I have not seen what it is at the time of a crash.

Any ideas please?
Philip

 

[circes9]

What has changed?

May want to check the versions of client/requesters being used.

 

[JockMullin]

What does the Novell server say about LAN problems?

I'm not familiar with version 6 but prev versions have monitor.nlm you can run on the console which allows you to display error counts for the network cards. Also check for timeouts and / or disconnects in the logs or on the console.
Look for log entries happening at about the same time as the network timeouts - maybe the server is in recovery mode and not responding.

Netware 6 defaults to IP transport - are you using that or have you enabled IPX? Are there any IP conflicts?

This smells like a network problem - are all the affected machines on the same network segment or attached via the same switch?

If nothing has changed in the config of the server and wkstns or in the clipper app then it is likely hardware or cabling. Is it only the clipper app having problems?

Jock

 

[HotEye]

Nothing has changed as far as I know.

I have swapped the DOS system out for now, then I found it had lost it's processor fan so it may have been heat. But it does not explain the 98's.

We have 3 main 16-port hubs forming the network, and about 3 small hubs off of 3 of the ports. IPX is installed.

The Clipper app is being enhanced still, but I do not think it is changes to that as only some workstations are affected.

There is a Monitor screen on the server, what numbers should I report? There have been a few error counters, but only a few. Server up time only 15 days. The server shows no signs of recovery mode and keeps running normally.

Only one Clipper app. in use.

Are there any settings to tweak in the clients? Should I upgrade to the latest?

 

[circes9]

Next time it happens, if you can get to the workstation quicky enough, it would be interesting to see if, from a command prompt you can do a directory of the server drive and "see" the file that the Clipper app couldn't find.

If the file is listed as part of the directory, but errors under Clipper, then it would suggest to look at the Clipper app. If the file does NOT appear as part of the directory, then don't look for a fix with Clipper because it is not at fault.

Does this DOS error 2 ONLY occur while running the Clipper app, or from other applications as well?

Are there alot of files in the directory Clipper is trying to read?

Good luck

 

[DTracy]

Is the filename using standard 8-3 convention or has someone accidentally created a long filename?



Best Regards,
David Tracy
dba: Tall Clipper Ships
Clifton, Arizona 85533
tracys@vtc.net

 

[HotEye]

Thanks for the luck.

The DOS systems only run the Clipper app. Usually a retry works so there would not be time to shell out and dir to see.

The W98 systems do only appear to error in the Clipper app that I am aware of. But that is the main app they run. Some run OExpress and Internet Explorer, but those appear ok.

There are about 600 files in the directory, including sub dir names. All files are 8.3.

Thanks for helping.

 

[JockMullin]

Hi, Hoteye

Regarding lan info in monitor.

As I say I have no experience with Novell 6, but in previous versions, on the main screen of the monitor there are a number of items you can select: Connections, Lan connections etc (terminology may vary, I don't have access to a Novell server here). Select the one for LAN and it will show you the or a list of the network card(s) in the server. Select one and hit enter and you will be shown a bunch of counters, some of which are error counters, which tend to be further down the list. Look for timeout errors or resets. If there is more than one nic, compare them & see if one has a significantly higher error count.

You could try posting in the Novell 6 forum - there are some very knoledgeable people over there.

You mention the stacked hub configuration of your lan. If they are hubs then it is possible for a fault in a workstation or cable to lock up the hub (and any downstream from it) and cause timeouts. Those will not necessarily be reflected in the server stats.

If you are having network problems then you have to establish where the fault lies, which may require an analyzer. However, since only some w/s's are affected, try moving those one by one to another hub and see if it still has problems or if the problem moves. Are the affected w/s's all on the same branch? If so that would be an indication it is a network problem.

Jock