Error Reading File

[pdchilders]

Client has a 5.0 system (won't be upgrading) that's really been running fine for years.

In the past week, has been receiving many errors scattered throughout modules.

Have found several files receiving "error reading file" errors when trying to reindex, open in foxpro, and/or copy via Win Explorer. These files invariably become available again 10-15 minutes later...One file that was "locked" was SYEELOG (tried to view in VFP no good, unable to even copy the file to backup, but was able to delete and copy in an empty version) and another ICYTRN01 (found locked during reindex, couldn't open in VFP, but 15mins later reindex IC ran fine).

Seems like some server/system process is locking files, but I'm not sure where/how to look. I had him turn off Norton Antivirus for now thinking it might be doing live scans and locking files...

I'm not onsite, and am not involved in server config/maintenance.

Thanks for any advice on this, he's really hurting at the moment...

 

[BrianoNur]

PdChilders,

"Error reading file" is an error message comes out when communication between worsktation and server is broken.

This interruption is temporarily. The question is why it happens now and why it doesn't before?

Ok, let's see a few of stuff, could be one of the following:

1. Have you increased the number of SBT 5.0 users?
Number of on-line users could hit the speed of response.
Remember VFP is not a client-server program, datafiles
takes longer to get the workstation memory. When users
run a report including history files, this message
probably come out when speed of response is low.

2. Have you changed the server from 32-bits to 64-bits?
SBT Pro 5.0 is a 32-bits software architecture.
Communication breaks down and that message comes out.

3. Have you checked out communication devices lately?
I have seen before that a misfunction on hub-switch,
server card, workstation netcard or cables could take
the communication down.

I think it is a good idea if you try to reindex files from the server, no workstations involved:

Here is in brief what you should do:
1. Create a SBT 5.0 shortcut on server desktop
2. Enter into SBT and get the System Manager.
3. Start running Index process, pick GL module only.

Open your eyes on speed, error messages, interruptions or whatever comes out.

How was it? If you got no error messages, no interruptions, focus on those 3 first points I mentioned above, problem could be one of them, otherwise problems could be on server memory or hard disks space.

Hope that helps you and you get the problem fixed soon.

 

[pdchilders]

All great things to try Brian, thanks!

Odd thing is both times I was troubleshooting after-hours (>8pm) when no users were even in the building (and no dial-up)...tried to open SYEELOG and another time ICYTRN01 and got same "error reading file" :-(

And, as said above, when I tried to make a file copy of the SYEELOG files via Windows Explorer, got "not enough memory" message, even though >27gb available on the drive...

Seems definitely server O/S issue, but I'm at a loss how to start looking...

Thanks

 

[BrianoNur]

Both datafiles SYEELOG and ICYTRN have high number of records, both have memo fields as well, that could explain memory error messages.

A tip for you is to delete old records from SYEELOG, leave current ones. It is a error log datafile, usually old records can be deleted since they are not useful anymore.

About ICYTRN datafile, there is not much to do, you should not delete any record from it.

If try a copy from hard disk to same hard disk, that error message (about memory) shouldn't come out. How do you do this?

 

[fitedgar]

I had a similar problem wit a customer. It turned out to be a back up software. My client had subscribed to a internet data back up Company. This company was backing up data (including SBT) every 2 hours. This created files not being read, index problems, transactions not posting, etc.

Edgar
Integrated Bar Code Systems, Inc.

 

[pdchilders]

So went onsite, they had new RAM for the server (were running 1gb, maxxed out the server RAM with new 2gb!), cold-boot (powered down to install RAM).

No more read errors since...

Not sure if it was the increased memory or the cold boot...

That network/server config is the least desirable scenario...

Thanks for the tips/head's up.


David

 

[BrianoNur]

David, Good for you!



Edgar,

I wouldn't never recommend to run a backup procedure every two hours, it could be a cause of certain problems.

Let's remember an open datafile can not be copied, unless backup program manages that. On the other hand, if a datafile is being copied it will not be open by SBT.

I would suggest to have a trustable server instead and make a backup process once a day.

Keep in touch.