Kernel upgrade to patch level 196, ECC 5.0, MSCS resource issue.

[Former Member]

Hi,

We have 3 system landscape DEV, QAS & PRD on win 2003, oracle 10G, ECC 5.0.

Our PRD system is on MSCS with CI on node A and DB on node B.

Recently I successfully upgrade kernel off DEV & QAS system to patch level 196 of stack 18.

When I tried to upgrade the kernel of PRD system one of the cluster resource drive (The drive contains the usr folder) is automatically moving from Node A to Node B and I got the error u201Capplication server shut downu201D. We donu2019t have additional application server in the landscape.

Please suggest the solution.

Regards,

Rajesh

[former_member308058]

Dear Rajesh,

If it is windows system, you may need to do a restart to clear virtual memory or if it doesn't solve ..

Try to change paging space to 0 and then reboot and then again to original size....

Regards,

akhilesh

[Former Member]

Hi Rajesh,

Take the resources offline using cluster admin. If possible down the another node and carry the activity.

I think the usr folder is on cluster resoures.

Regards,

Rahul.

[Former Member]

Rahul,

Yes, usr folder resides on cluster resource.

I checked the dev_w0 file & found the below error;

ERROR => [MapOsShm] Can't find free space for Shared Memory (Size=83582 KB [shmnt.c 2130]

ERROR => [AttachOsShm] MapViewOfFile(Key=62,Handle=0x000004A4) failed with Err=0 [shmnt.c 2130]

ERROR => ShmCreate: Attach (62,85588608,2) failed [shmnt.c 385]

Thanks,

Rajesh

[JPReyes]

Can't find free space for Shared Memory 

Pretty clear message... what version of SAP are you on?... you might be able to use the cleanipc XX command and start the system again.

Regards

Juan

[Former Member]

Hi Juan,

How are you? Thanks for your reply.

We are using ECC 5.0. I have found more errors in dev_w0 as follows,

• ERROR => ShmCreate( ATTACH 62, 85588600) failed 1 [mpixx.c 843]

• ERROR => DpInitMpi: MpiInit failed (1) [dpxxtool2.c 3556]

in_ThErrHandle: 1

• ERROR => tskh_init: DpInitMpi (step 1, th_errno 17, action 3, level 1) [thxxhead.c 9621]

Info for wp 0

stat = 4

reqtype = 1

act_reqtype = -1

rq_info = 0

tid = -1

mode = 255

len = -1

rq_id = 65535

rq_source = 255

last_tid = 0

last_mode = 0

int_checked_resource(RFC) = 0

ext_checked_resource(RFC) = 0

int_checked_resource(HTTP) = 0

ext_checked_resource(HTTP) = 0

report = > <

action = 0

tab_name = > <

*****************************************************************************

*

• LOCATION SAP-Server sapcigrp_ECP_00 on host sapeccprddb (wp 0)

• ERROR tskh_init: DpInitMpi

*

• TIME Sun Nov 16 05:13:07 2008

• RELEASE 640

• COMPONENT Taskhandler

• VERSION 1

• RC 17

• MODULE thxxhead.c

• LINE 9806

• COUNTER 1

*

*****************************************************************************

PfStatDisconnect: disconnect statistics

Entering TH_CALLHOOKS

ThCallHooks: call hook >SAP-Trace buffer write< for event BEFORE_DUMP

TrThHookFunc: called for WP dump

ThCallHooks: call hook >ThrSaveSPAFields< for event BEFORE_DUMP

• ERROR => ThrSaveSPAFields: no valid thr_wpadm [thxxrun1.c 730]

• ERROR => ThCallHooks: event handler ThrSaveSPAFields for event BEFORE_DUMP failed hxxtool3.c 254]

Entering ThSetStatError

Entering ThReadDetachMode

call ThrShutDown (1)...

***LOG Q02=> wp_halt, WPStop (Workproc 0 4000) [dpnttool.c 357]

Please suggest the solution

Thanks & Regards,

Rajesh

[JPReyes]

Seems like you need to clear the shared memory before been able to start the system.... In older sap versions you can use kernel command cleanipc XX (where XX is the instance number) to clear the shared memory... In the new releases is different tho... theres plenty of threads about it here at SDN... a quick search will give you a few clues.

regards

Juan