Work processes going in ENQ and system is getting slow

[former_member280828]

Hey Experts,

Hope everyone is doing great,

I have noticed that once in every month whenever i checked my SM66 i see my work processes going in ENQ and my system is getting slow.

I always try to kill the long running ENQ processes but nothing happens system doesn't come in normal state and for this reason i always need to take a downtime and restart the system again.

We are having ESS MSS running in our system with 3 R3 Application server and 5 Java application server.

I have checked the lock entries there are only 38 lock entries this time when i checked.

Can anyone please help me analysing this issue.

I want to know why this happens once in every month ? and what can be done to avoid this situation ? and if this situation arises what can be done to make system normal instead of restarting the SAP.

Please help me.

Thanks and regards,

Musa.

[feng_shi · Accepted solution]

Hi Musa,

SM66:

workprocess Running ENQ SAPLSENA

which means that the workprocess has sent enqueue request to ENQ server, but there is no response from ENQ server yet.

I suppose that you are using ASCS standalone enqueue server.

the possible causes are:

1, ENQ server problem

=> if so, you should find some errors from dev_enq* trace file

2, network issues which block communication between DI and ENQ server

=> you can ask you network admin to troubleshooting it when issue is happening again

3, are you enable ERS server ? it could be synchronize problem between ENQ server and ERS server

=> you can just try to disable ERS server and check again

Since this issue happens once in every month, does this issue occur on the same period every time ?

what did you perform during that period every month ?

Probably the special actions which performed every month cause this issue(see possible reasons above).

Thanks and best regards,

Shi

[former_member280828 · Accepted solution]

Hey Feng,

How are you?

1) Where can i find the dev_enq* trace file ??

2) I dont have network related problem.   

3) No ERS server.

I analysed that this issue arises in month end payroll processing activity.

Becuase at payroll processing HR's uses Tcodes and process the reports in Dialog WP which possibly runs for 50 mins or 60 mins or sometimes more than that and that time system also runs very slow and gives performance issues.

Regards,

Musa.

[Former Member · Accepted solution]

Hi Musa,

If you have ASCS configured, then the location would be

/usr/sap/SID/ASCSxx/work directory or else /usr/sap/SID/DV***xx/work directory.

Also,

SM12 -> “test” in OK/command window then you will see extra option in menu bar Error handling -> Diagnosis Environment

will give the complete diagnostic report over locking mechanism. I have tried many a times in my landscape without any problem. Don't why such an option from SAP would harm the system?

Regards,

Prithviraj.

[former_member280828 · Accepted solution]

Hey Prithviraj,

What i wanna ask you is, at present my work processes are working fine because before payroll when i had that problem then i took a downtime and then restarted SAP system.

So now do i have to perform all these steps to find the solution now or at the time when the ENQ issue comes ???

Regards,

Musa

[Former Member · Accepted solution]

Hi Musa,

For steps mentioned below, you can perform anytime to test the robustness of locking mechanism.

SM12 -> Goto/Extra -> Diagnosis/Diagnosis in Update and

SM12 -> “test” in OK/command window then you will see extra option in menu bar Error handling -> Diagnosis Environment

If both the reports are clean, than the issue is not on SAP locks (SM12) side.

For trace/logging i.e. Error handling - Logging - Switch On, you need to perform at the time of issue. make sure the trace should not be longer than 10-15 secs.

You can try to understand the logs using sap note 125041

Regards,

Prithviraj.

[former_member280828 · Accepted solution]

Hi Prithviraj,

SM12 -> Goto/Extra -> Diagnosis - This step gives the following output

Error: Local lock management process despite external lock management

Error: System contains more than one lock management CPCAPP110_SBP_01

Test lock operations do not contain errors.

Optimistic locks are implemented in the kernel

Test lock operations do not contain errors.

Test lock operations do not contain errors.

Regards,

Musa

[Former Member · Accepted solution]

Hi Musa,

The error reported in the diagnosis points to having more than one ENQ work process? Please check.

Error: Local lock management process despite external lock management

Error: System contains more than one lock management

39412 - How many work processes should be configured?

Also check if you have rdisp/enqname is configured correctly.

Regards,

Prithviraj.

[former_member280828]

Sorry Guys for the very late reply. Actually i was very busy in Year End activities.

Thanks everyone for their valuable reply in my absence.

I have followed couple of SAP notes but according to that notes i couldn't be able to resolve my issue.

http://service.sap.com/sap/support/notes/1384191

I followed this above note

Solution -

• Call transaction SM12. In the command field (also called OK code field), enter "test" (without the "") and press the Enter key.
• After you do this, the system displays another pull-down menu called "Error handling".
• From the pull-down menu, choose "Error handling - Logging - Switch On" and wait for approximately ten seconds.
• Then choose "Error handling - Logging - Switch Off" from the pull-down menu.
• Finally, choose "Error handling - Logging - Display" from the pull-down menu.
• Save the list that is displayed in a local file. When you do this, choose "unconverted" for the format in which the list should be saved. You can compress the file and attach it to a message.

I followed these above given steps but the logs are coming empty after performing these steps in SM12

Regards,

Musa

[isaias_freitas]

Hello Musa,

Does the affected system have an ASCS instance?

If yes, the enqueue will be running there.

The WIKI page I sent earlier has the steps to activate/deactivate the enqueue logging at the standalone enqueue (at the ASCS).

Notice that this log is just for you to identify what application is overloading the enqueue with requests. After the application has been identified (either by the analysis of the log, or if you already know which is the application - Year End activities / jobs?), you would need to look for a solution from that application area.

Regards,

Isaías

[raquel_gomez]

Hi,

You mentioned that this issue, with work processes goind in ENQ and system getting slow only happens during certain periods; therefore this issue seems to be related to isolated moments with high Enqueue load.

The report SAPLSENA should not be reported in transaction sm50/sm66, as this response is given quite quickly.

Having lots of WP running with SAPLSENA report can happen due to two reasons:

- The application tries to set a large number of locks.

- The requested lock may not be set due to a lock collision. The application repeatedly requests the lock (busy loop) until a lock collision no longer occurs.

In this case, problem seems to be related to same application trying to set many locks. Note 1384191 - SM50: Many processes in status "holds ENQ"contains the explanation on how to create an enqueue log. It will help us to identify the application.

Also check following error from SM12 > Goto > Extras > Diagnosis:

Error: Local lock management process despite external lock management

Error: System contains more than one lock management

These errors indicates that you have more than one enqueue process configured in your system.

Is the system running with Standalone Enqueue Server (ENSA)? In that case this will provide the enqueue service, so you do NOT need an enqueue work process in the central instance anymore.

Regards,

Raquel

[former_member280828]

[former_member280828]

Hi Raquel,

See the Landscape Picture i have uploaded.

I am having a HCM, ESS MSS running on my landscape with approx. 11000 active users.

So everytime i see sm12, I see a Huge amount of locks in report.

Regards,

Musaddik.

[raquel_gomez]

Hi,

There's no problem on having many locks on SM12, if this is needed by the system, as long as there are enough resources (enqueue table size is big enough) and those locks are released afterwards correctly by the applications.

You complained about having work processes going in ENQ and system getting slow. As already mentioned, report SAPLSENA should not be seen in transaction sm50/sm66.

This may happen if an  application tries to set a large number of locks; or the requested lock may not be set due to a lock collision. You can create an enqueue log (as per Note 1384191) to identify which is the involved application causing this situation.

Also ensure that  there's only one Enqueue Service configured on the system. If system has defined Standalone Enqueue Server (ENSA); there's no need to maintain ENQ work process on Central Instance.

Regards,

Raquel

[former_member280828]

Hey Raquel,

I followed the notes 1384191 but after switching the logging on and off theres nothing coming in logs as per the notes.

I tried it earlier also but logs comes blank.

What should i do.

Regards,

Musa.

[raquel_gomez]

Hi Musa,

You should firslty ensure that Enque Server is running just one. Please check output from
transaction SM12 , menu Extras > Diagnosis. Please provide an screenshot from the output.

Also note that Enqueue log needs to be taken when the problem is happening (work processes in status ENQ showing SAPLSENA).

Regards,
Raquel

[Former Member]

Hi musa,

I'm facing same situation.

Could you please tell me how you resolved this issue ?

- situation

All dialog workprocesses are occupied by report SAPLSENA.

This problem happened at a day after monthly OS maintenance.

Regards,

Ishii

[avadhesh_sap]

Hi Ishii,

Please check if enqueue service is running at os-level.

Use command ensmon for linux/unix.

Thanks,

Avadhesh

[Former Member]

if it's on BALHANDLE

can you try if these notes gets applied to your system

http://service.sap.com/sap/support/notes/2025948

http://service.sap.com/sap/support/notes/1998968

http://service.sap.com/sap/support/notes/2069753

I had same issue in Solman and the above notes resolved them.

[raviraj_sap]

Hi

Double click on the ERROR column & see what error you are getting ?

Also, check dev_enque* trace file for any issues.

Thanks

Ravi

[Former Member]

Hi,

Perform diagnostic test as per below steps and check the report generated.

SM12 -> Goto/Extra -> Diagnosis/Diagnosis in Update and

SM12 -> “test” in OK/command window then you will see extra option in menu bar Error handling -> Diagnosis Environment.

As per note mentioned by Yuksel trace the enqueue for 10 sec and with the help of the below note you might be able to understand it

125041 - Analysis of enqueue errors with enqueue logging.

Check if the parameters are set as the link for ASCS instance

Profile Parameters for the Enqueue Clients - Standalone Enqueue Server - SAP Library

Check the below sap notes and wiki

2069753 - Infinite wait in program SAPLSENA

43614 - Lock entries remain after end of session

FAQ Enqueue Server - Application Server Infrastructure - SCN Wiki

Regards,

Prithviraj.

[isaias_freitas]

Hello,

SM12 -> “test” in OK/command window then you will see extra option in menu bar Error handling -> Diagnosis Environment.

Be VERY CAREFUL with that . Some menu options under "Error handling" can break the system.

Another possibility for the work processes to be waiting for the enqueue is that an application is overloading the enqueue with requests. You have mentioned that you see this once per month. Is it during the same period of the month? Let's say, during moth end closing?

What specific jobs are executed then?

About the note 1384191, check out this WIKI page as well.

Cheers!

Isaías

[yakcinar]

Hello Rbk,

Can you check the note 1384191 - SM50: Many processes in status "holds ENQ"

And the link https://scn.sap.com/thread/1658110.

Regards,

Yuksel AKCINAR