cancel
Showing results for 
Search instead for 
Did you mean: 

Work Processes Stucked and Lock table error

Former Member
0 Kudos

Hello Expert,

We are using ECC 5.0 640_EX2 Kernel and patch level 304.

For Production server we are using Cluster HA scenario.

We have DB CI in cluster and 18 apps server connected to CI.

We faced following problem repetative time on this server

Problem : Not able to login CI server.

Observation : 1. All DIAG WPs are stuck on CI server

DIAG WPS in running state but nothing is running actually (No reports nothing shown)we tried to kill few DIAG WPs but not released. We monitored this at OS level (AIX 5.3) using DPMON.

2. Lock error found in SM21 log

There was no more activity possible on server as no more lock possible and lock error occuring.

Issue resolved after restart of the CI server and manually clean of IPCs processes.

Kindly let us know where and how can we check the reason for such incident and precautions we can take?

Is this any know issue? and what can be the reason?

Regards,

Nitin Salunkhe

Accepted Solutions (1)

Accepted Solutions (1)

former_member524429
Active Contributor
0 Kudos

Hi,

Which Error message you were getting in SM21 log. Post the relevant log entries.

Remove Old Lock Entries using SM13, if they are there.

CI and All Application Servers are having same SAP Kernel Patch Level ?

If possible you can apply latest SAP Kernel Patch, disp+work package, lib_dbsl package, tp, R3trans. Maintain same patch level on all instance.

The existing size of Enque table is enough to handle lock operations in pick load ?

check parameter enque/table_size. If its required enlarge it after analyzing Diagnostic information of Enque Process (SM13). If Lock table is overflow then u will find Lock table Overflow error message in SM21 log.

Regards,

Bhaivk G. Shroff

Answers (1)

Answers (1)

Former Member
0 Kudos

You need to first trace why your lock table entries filled up. There may be something went wrong due to your lock table got full. Check for any background job or any specific user has triggered something in your system which should not be done. If you check in SM21 and trace what are the users affected? There are few reports in SAP which will overflow your lock table (ie mass changes).

Former Member
0 Kudos

Hello,

Thanks bhavik and sunil for reply.

I am getting following lock error in the system. Also I found following errors in SM21

Transaction Canceled 14 025 ( )

Transaction Canceled 14 025 ( )

GI0 Error calling the central lock handler

GI2 > Unable to reach central lock handler

D01 Transaction Canceled 14 025 ( )

D01 Transaction Canceled 14 025 ( )

When there was a problem I observed that all DIAG wps full but only showing process ID and nothing was running actually.

I think DIAG WPs was not getting released do you have any idea about reason behind this kind of behaviour.

because I doubt lock table get full it has sufficient size peak value is not reached to high

Regards,

Nitin Salunkhe

Former Member
0 Kudos

Is the Note 734728 - Syslog GI0 and GI2 (unable to reach central lock handler) valid for you?

former_member524429
Active Contributor
0 Kudos

Hi,

Along with the above suggestion, Have you referred [SAP Note 488036 - Unable to reach central lock handler, AIX only|https://service.sap.com/sap/support/notes/488036] ?

Such error message may come due to Networking issue. Also consult your OS administrator to look at TCP/IP send/receiver buffer parameter at OS level (like tcp_sendspace, tcp_recvspace in AIX). Try to enlarge them if they are set with small values.

Regards,

Bhavik G. Shroff