MtxLock DeadLock problem

[former_member634145]

Hello all,

In one of our ECC system when a user executing one transaction work process is taking more time for execution. When i checked the trace file of the work process i found following error/warning:

I WARNING: MtxLock 0xc000000017e340f0 rrol0003 owner=2 deadlock ?

I Fri Mar 11 16:18:43 2011

I WARNING: MtxLock 0xc000000017e340f0 rrol0003 owner=2 deadlock ?

I Fri Mar 11 16:19:44 2011

I WARNING: MtxLock 0xc000000017e340f0 rrol0003 owner=2 deadlock ?

I searched on google but could not find anything. On SDN i found only one forum and some one has told to check if note 1309072 - System standstill: MtxLock EMADM helps to resolve the issue. But this note also could not help.

Can anyone please answer this what this is exactly and why this occurs in system???

Regards,

Nirav

[Former Member]

Hi Nirav,

Please give full trace of the respective work process (dev_w*) to check if we something more is found to

detect the issue in particular.

Regards,

Sujit.

[Former Member]

Hi,

Can you please let me know in detail to your corrections of below?

1. did you upgrade your Kernel?, what was your Kernel before and after?

2. did you also change and check the parameter "rdisp/check_master_slave_active" to value '2'?

OR you have opted Workaround of the note:1309072?

Regards,

Sreenivas .Y

[former_member634145]

Hello,

Yes i have already check note 1309072 and already set parameter rdisp/check_master_slave_active" to value '2'.

Regards,

Nirav

[Former Member]

Was the parameter set in all servers/instances or at default profile?

Regards,

Sreenivas .Y

[former_member634145]

I set it in default profile.

Regards,

Nirav

[Former Member]

Nirav,

Is your kernel patch is latest one?

Regards,

Sujit.

[former_member634145]

Hello Sujit,

My Kernel is at 111 patch level.

Regards,

Nirav

[Former Member]

Hi Nirav,

Alright. So, its an updated one.

Please give us the dev_w* trace. I already requested you earlier also.

Let us check if we can find something in more detail.

Regards,

Sujit.

[Former Member]

Yes dev trace may hlep much, did you restart your SAP after the parameter set and cross checked in all instances?

Regards,

Sreenivas .Y

[Former Member]

I'm facing the same problem too and it occurred several times before, especially when someone was running programs which cost too much time and decided to ternimate the session but for some unknown reason failed, then the GUI crashed and after that, the work process began deadlock.

The trace log is like below:

C Thu Apr 14 15:50:56 2011

C build_stmt: reallocating stmt buffer: 3801 -> 4807 characters

C build_stmt: reallocating stmt buffer: 4807 -> 5808 characters

C build_stmt: reallocating stmt buffer: 2304 -> 3308 characters

C build_stmt: reallocating stmt buffer: 3308 -> 4318 characters

C build_stmt: reallocating stmt buffer: 512 -> 2000 characters

C build_stmt: reallocating stmt buffer: 3072 -> 4078 characters

C build_stmt: reallocating stmt buffer: 4078 -> 5090 characters

C build_stmt: reallocating stmt buffer: 256 -> 2000 characters

I

I Thu Apr 14 16:09:07 2011

I WARNING: MtxLock 0x7000000063626e4 rrol0112 owner=0 deadlock ?

I

I Thu Apr 14 16:10:08 2011

I WARNING: MtxLock 0x7000000063626e4 rrol0112 owner=0 deadlock ?

I

I Thu Apr 14 16:11:08 2011

I WARNING: MtxLock 0x7000000063626e4 rrol0112 owner=0 deadlock ?

We are using ECC6 with kernal SAPKB70103 on AIX system with oracle database. I also noticed sap notes 1309072, but i don't know what EMADM stands for in the note as we have 'rrol0112' instead of 'EMADM' in the log, and whether we should apply the patch SP199 before let it work or we can just set parameters as mentioned in the notes.

Could anybody please help to explain it or tell me how to do with the deadlock?

Thank you very much!

Best Regards,

Jeff