cancel
Showing results for 
Search instead for 
Did you mean: 

ENQ processes and semaphore 26 in SM66/SM50, SAPLSENA & SAPLSENT

Former Member
0 Kudos

Hi

We quite often see a number of DIA and ENQ processes in status 'running' with a Semaphore 26 error in SM66/SM50. Most of these (but not always all) show reads on tables SAPLSENA and SAPLSENT.

Many other processes - DIA, BTC and UPD - then have the 'stopped ENQ' status and also show reads on tables SAPLSENA, SAPLSENT and SAPLTHFB.

This doesn't last for that long but seems to be a common occurrence and we often get performance complaints during this time.

Note 33873 - What do the semaphores mean? - shows that this is the ENQ-TABLE-Semaphore.

One suggestion was that a filesystem backup running at the time was locking a file (enque related file??) but we have not been able to substaniate this. We could run the command fuser but which file is being locked?

We have already followed SAP Note 1384191 - Transaction SM50: Many processes with report SAPLSENA, but could not see any useful indicators in the log file produced.

Has anyone else experienced this or know how to check what is causing this?

Thanks

Ross

Accepted Solutions (1)

Accepted Solutions (1)

Former Member
0 Kudos

Hi Ross,

In case you are using a backup file for your enqueue, you should check parameter enque/backup_file

Are you using a standalone enqueue, or is the enqueue part of the central instance?

Do you see any errors in dev_enqwork

Kind regards,

Mark Dijsselbloem

Former Member
0 Kudos

Hi Mark

enque/backup_file is set to the default, /usr/sap/S1P/DVEBMGS00/log/ENQBCK.

We have already ran fuser on this file, and although there appear to be a LOT of processes accessing the file, none of these belong to the filesystem backup which we thought may be locking it.

We are not using a standalone enqueue, the enqueue is part of the central instance.

What do you mean by dev_enqwork? I checked the dev_* files in /usr/sap/S1P/DVEBMGS50/work but there are a lot of files and a lot of meanlingless information, can't see anything useful here...

Are there any other files that the filesystem backup could be locking that could cause these issues? It seems to be backing up files in /usr/sap/S1P ... not sure it is this though as I have just seen ENQ semaphore 26 processes and there is no backup running now....

Ross

Former Member
0 Kudos

Hi Ross,

With dev_enqwork I mean the developer trace of the enqueue server. If you have problems with your enqueue, I would expect some messages in there.

Have you also checked the kernel parameters of your system? It could be that you exceed the number of open files or sessions.. If there is a problem related to this, you should see that in the system log of the operating system.

To check if the problem is related to the backup file for the enqueue, you could disable the backup file for a while. This might bring some new insight into this issue...

Kind regards,

Mark Dijsselbloem

Former Member
0 Kudos

The problem is not presenting itself at the moment, but when it does again I will increase the trace level of the process and check the relevant dev_w* file.

Our UNIX team haven't advised us of any such errors but we could check this.

We are also looking into moving backup times.

Thanks for your help

Ross

Answers (0)