Hello friends
I am still using TPS, my system is large and the database project is ongoing.
I have the following problem, which started this month with just one client.
The problem:
Only one table (tps), when the number of accesses (network users) increases, remains allocated, uncorrupted. At this time, the table cannot save any data, the other users who are on a screen that contains the table, freeze and the client has to close the system on some computers and the table is released.
I have already looked at all the code related to access to this table and there is no Logout() without commit, or Stream() without Flush().
The table is saved in only 2 procedures
1 - The user opens the procedure and a list of records related to a request is saved in this table
2 - An exe saves the same list of requests, every 40 seconds if there is data for these records.
This has been used for a long time, years and by several clients, but only 1 started reporting this problem this month.
I have already done:
-I understood that the table is allocated because at the time of the problem I opened the table through topScan, changed some data and tried to save it at that moment the table froze. We did the procedure of closing the system on the computers, one by one, and suddenly it unfroze and let me save.
-I changed the recording sequence
-I ran topsFix on the table with and without .tpe
But the problem continues if a large number of users start using it.
The entire system works normally, the problem only appears when this table is requested.
It is confusing, but that’s it.
I will try to change the HD on the client, because I don’t know where else to go.
Any extra info when you turn on logging?
Are you checking for errors?
How many records in the table?
How big is the file?
Any unique keys? If so, does a blank record exist?
Hi Jslarve,
Any extra info when you turn on logging? → Not at first
Are you checking for errors? → I didn’t return the error, but it seems to be File Is Already Locked although I didn’t put the LogOut() on this table
How many records are there in the table? → 873896
How big is the file? → 116175 Kb (117MB)
Any unique keys? If so, is there a blank record? → Yes have a unique key… I ran a tpsFix rebuilding the keys and got no errors… I believe there are no blank records from what I analyzed.
Have you got any antivirus software monitoring the file, you can tell AV software to ignores some files and folders which can speed things up if it’s monitoring without any errors.
This util can tell you what file handles exist.
The number of open file handles is 16,711,680 on 64 bit machines and on 32 bit machines it will be 16,744,448. That might be a factor if something is not closing file handles.
Is there any shadow copy enabled on the folder. This is used by backup software and can cause problems if it’s backing up too frequently ie one backup job running and another one queued to run immediately after.
What OS is it, what AV software and what backup software?
is completly consistant with a lock, which a logout also does. Note that the ABC classes also use logout on things like a relational update. So while you may not have added it, it might still be there.
This might also be active on an open normal form that someone is in the middle of updating.
You could try turning off the logout in the templates, thats a checkbox somewhere, and see if that helps
Hello everyone, I found a LogOut without the commit, on a screen where the commit was in an embed that was commented (!commit).
Distraction when deactivating an embed.
The screen in question is used occasionally so I didn’t pay attention to it.
Thank you very much for everyone’s attention and time