SAPOsCol running but not working (shared memory not available)

[Former Member]

Dear Forum,

We have just set new passwords for SAP AD users (incl SAPService<SID> and <SID>adm), after this restarted SAP instances and servers / clusters). Everything came up well and all services, incl SAPOsCol started up automatically with the new passwords.

Everything works fine - only this morning i wanted to have a look in ST06 and the program tells me "SAPOSCOL not running? (shared memory not available).

The service is running tho, would a restart of the service do any difference, and is there any possibility that a restart of the service on a live and running system would force a system restart? (i couldnt think of any, but I hate restarting services on a live running production system).

Would it be an option to stop/start the collector from ST06?

Thanks in advance,

Kind regards,

Soren

[Former Member · Accepted solution]

Hi,

You can restart the SAPOSCOL service on live running system.

from ST06 -> click on operating system collector here yopu found option for Start , Stop service.

regards,

kaushal

[Former Member · Accepted solution]

Hello Kaushal,

Thank you for your answer!

I tried to start and stop/start the service from ST06, but it doesnt work. Here is a content og the log - any ideas how i could start the collector without booting the server?

HWO description

SAPOSCOL version COLL 20.95 701 - 20.64 NT 07/10/17, 64 bit, multithreaded, Non-Unicode

compiled at Feb 24 2009

systemid 562 (PC with Windows NT)

relno 7010

patch text COLL 20.95 701 - 20.64 NT 07/10/17

patchno 32

intno 20020600

running on L5183N01 Windows NT 6.0 6002 Service Pack 2 16x AMD64 Level 6 (Mod 26 Step 5)

PATCHES

DATE CHANGELIST PLATFORM PATCHTEXT

20081211 1032251 ALL Option -w support. Removed SizeOfRecord warning rt 130.

20090114 1036522 UNIX Log file permissions: 664.

20090203 1041526 UNIX Add. single instance check.

20090210 1042962 ALL Continue after EWA HW XML generation failure.

09:46:39 12.04.2010 LOG: Profile : no profile used

09:46:39 12.04.2010 LOG: Saposcol Version : [COLL 20.95 701 - 20.64 NT 07/10/17]

09:46:39 12.04.2010 LOG: Working directory : C:\usr\sap\PRFCLOG

09:46:39 12.04.2010 LOG: Allocate Counter Buffer [10000 Bytes]

09:46:39 12.04.2010 LOG: Allocate Instance Buffer [10000 Bytes]

09:46:40 12.04.2010 LOG: Shared Memory Size: 118220.

09:46:40 12.04.2010 LOG: Connected to existing shared memory.

09:46:40 12.04.2010 LOG: MaxRecords = 1037 <> RecordCnt + Dta_offset = 1051 + 61

09:46:55 12.04.2010 WARNING: WaitFree: could not set new shared memory status after 15 sec

09:46:55 12.04.2010 WARNING: Cannot create Shared Memory

thanks!

Soren

Edited by: Soeren Friis Pedersen on Apr 12, 2010 9:49 AM

[Former Member · Accepted solution]

Hi.

try to do following at os level

1. saposcol -d ; kill ; leave ; launch ; exit.

2. start saposcol from st06

under which os account saposcol was running? also check windows event log.

share your result.

regards,

kaushal

Edited by: Kaushal Malavia on Apr 12, 2010 1:35 PM

[Former Member · Accepted solution]

Hi kaushal,

the service is running with user <SID>adm. Is there any risk of performing the commands in line 1? is the syntax as presented or is kill leave launch exit in separate commands?

Thanks in advance! I really appreciate your kind help.

[Former Member · Accepted solution]

Hi,

I found one note for your problem. see the topic # 7.

[Note 548699 - FAQ: OS collector SAPOSCOL|https://service.sap.com/sap/support/notes/548699]

You may also need to increase PHY_MEMSIZE profile parameter in RZ10

YOU CAN ALSO TRY USING LATEST "SAPOSCOL".

It is always recommended....

Have a look at below note also.

[Note 19227 - Open newest saposcol|https://service.sap.com/sap/support/notes/19227]

Regards.

Rajesh Narkhede

[Former Member · Accepted solution]

Hi,

There is no risk to execute saposcol -d .

check http://help.sap.com/saphelp_nw04s/helpdata/en/c4/3a6c40505211d189550000e829fbbd/content.htm

for Control SAPOSCOL from the Operating System

also check SAP note suggested by Rajesh.

Regards,

kaushal

[Former Member · Accepted solution]

Hello Guys,

Thanks a lot for your help!

I tried from OS level (i had a lot of saposcol files on the system, but i took the one that the path of the service was pointing at). It didnt seem to work, I'm afraid i cant solve this without a system boot.

Does any of this help?:

Collector > kill

kill

14:21:06 12.04.2010 LOG: check_remap: Switched to new shared memory.

PID 0 not found. Collecting process not running ?

Setting Stop Flag :

14:21:21 12.04.2010 WARNING: WaitFree: could not set new shared memory status af

ter 15 sec

14:21:21 12.04.2010 LOG: ==== Stop Flag was set by saposcol (kill_collector())

.

14:21:21 12.04.2010 LOG: ==== The collection process will stop as soon as pos

sible

Collecting process has stopped.

Process died without reaction ?

14:21:32 12.04.2010 LOG: Shared memory deleted.

Collector > launch

Now connected to shared memory.

launch

14:22:00 12.04.2010 LOG: check_remap: Switched to new shared memory.

Cannot create Shared Memory

Collector > kill

Now connected to shared memory.

kill

PID 0 not found. Collecting process not running ?

Setting Stop Flag :

Collecting process has stopped.

Process died without reaction ?

Collector > leave

Now connected to shared memory.

leave

Shared memory deleted.

Collector > launch

Now connected to shared memory.

launch

Cannot create Shared Memory

Collector >

[Former Member · Accepted solution]

Hi,

I suggest that the "cold" reboot of server solves the shared memory problem , since it is your production server you need downtime.

regards,

kaushal

[Former Member · Accepted solution]

Hi Kaushal,

I was hoping it was possible without a restart of the system, I had it down just yesterday for AD pw changes. Oh well. thanks a lot for your help in this matter, it has been most helpful, and thanks again for spending your time to help solve my problem.

Kind regards,

Soren