livecache issue LC10 BY0 > Too many users (task limit)

[Former Member]

Hi,

we have 2 application servers with one CI/DB, maxusertasks is 500

we figured out a livecache pb, with the following error messages

(db analyzer)

LC10 BY0 > Too many users (task limit)

User task 281 blocked in stae 'Prep-End(230)' since 902s, DB procedure : SAPAPO_CLEANUP application pid

-

---

User task 281 blocked in stae 'Prep-End(230)' since 902s, DB procedure : SIM_SIMSESSION_CONTROL, pid 31685

...User task 281 blocked in stae 'Prep-End(230)' since 902s, DB procedure : APS_ACT_SCHEDULE, pid 3168

before we had the warining: UTK5 is running 1804s, CPU 50% and CPU 100% (LC)

we have checked knldiag/knldiagerr files and we are unable to find what is the root cause of the problem?

How could we analyse this issue and find the treatment which triggers this issue?

(trace: (SAPAPO/OM01)

Thanks and Best regards,

any help would be rewarded.

Best regards,

John

Edited by: DCWE COREADMIN on Feb 11, 2011 6:54 PM

[former_member229109 · Accepted solution]

Hello John,

We need additional information about the liveCache version, OS of the liveCache server; more details about situation when the issue occurred. Run "x_cons <SID> sh all 10 10 > task_info.txt" when the issue occurred to collect more information.

Check if the savepoint task is active - "x_cons <SID> show active".

Is the issue occurred during high loading to te liveCache?

Please review the SAP note 1391322 where the SPORADIC standstill in high load situation is also described, this issue is already fixed in the new liveCache releases.

This situation occured sporadically. It will be helpful to login to your system for further analysis of the reported issue to get clear if your system run into this known problem because the savepoint was hanging, because of that you had big amount of the user tasks which were hours in the state 'Prep-End' until the liveCache restart.

< It could be the reason for "livecache issue LC10 BY0 > Too many users (task limit)", you reached the MAXUSERTASKS limit,

the increasing of the liveCache parameter will not help you to solve the issue. >

You checked/saw the Database Analyzer log :

"User task 281 blocked in stae 'Prep-End(230)' since 902s, DB procedure : SAPAPO_CLEANUP application pid

...

User task 281 blocked in stae 'Prep-End(230)' since 902s, DB procedure : SIM_SIMSESSION_CONTROL, pid 31685

...User task 281 blocked in stae 'Prep-End(230)' since 902s, DB procedure : APS_ACT_SCHEDULE, pid 3168

"

You are SAP customer. As already recommended by Ivan, create the SAP message to the component "BC-DB-LVC" and get SAP support.

Thank you and best regards, Natalia Khlopina

[Former Member · Accepted solution]

Hello Ivan, Natalia,

Livecache Version is 7.6.02 BUILD 014-123-152-175

Solaris 10 5/09 s10x_u7wos_08 X86 (we have 2 application server on x86_64 GNU/Linux - redhat 5.5)

SPS 10 SCM 5.0

(we are very late..., but it planned to upgfrade in the next month)

Thios problem occured during a high downlaod ed liveache (CPU 100%), but the problem is we don't find which process cause the bottleneck (PPDS or DP/SNP?), how do you knwo how to relate the UTK to the user/livecache activity?

a SAP customer has been also opened.

Thanks for your help , Best regards

John

[Former Member · Accepted solution]

Hello Ivan, Natalia,

Livecache Version is 7.6.02 BUILD 014-123-152-175

Solaris 10 5/09 s10x_u7wos_08 X86 (we have 2 application server on x86_64 GNU/Linux - redhat 5.5)

SPS 10 SCM 5.0 (the note does not match with our version but thanks)

(we are very late..., but it planned to upgrade in the next month)

Thios problem occured during a high download on liveache (CPU 100%), but the problem is we don't find which process cause the bottleneck (PPDS or DP/SNP?), how do you knwo how to relate the UTK to the user/livecache activity?

a SAP customer has been also opened.

Thanks for your help , Best regards

John

Edited by: DCWE COREADMIN on Feb 15, 2011 10:31 AM

[former_member229109 · Accepted solution]

Hello John,

Please also review the SAP Note No. 1112046, if you run the liveCache on Solaris 10 Update less as 5.

What do you mean "a SAP customer has been also opened"?

Did you open the SAP message, where this issue was reported?

Regards, Natalia Khlopina

[Former Member · Accepted solution]

Hello Natalia,

Thanks for your help. we appreciazte.

we 're in solaris 10 update 7

Solaris 10 5/09 s10x_u7wos_08 X86

yes a sap message has been opened.

(I will send you the number if you want)

Best regards,

John

[former_member229109 · Accepted solution]

Hello John,

was your issue solved?

Regards, Natalia Khlopina

[Former Member · Accepted solution]

Hi Natalia,

unfortunately, the issue was not solved, but we haven't had this problem since last time... (keep fingers crossed)

we have checked all the parameters and we have upgraded support packga to Level SP22

(sapapo KERNEL 7.6.06 BUILD 011-123-234-767)

anyway, we still have huge response time (RFC time roll wait time) for pegging, scheduling table.

(I will open a new thread for this)

Regards,

John

Edited by: DCWE COREADMIN on Mar 7, 2011 10:53 AM

[Former Member]

Hi ,

you forgot to mentioend

->What is liveCache version / build ?

->What operating system do you use ?

Also I would suggest to open message with SAP to check it

Regards

Ivan

[markus_doehr2]

> LC10 BY0 > Too many users (task limit)

> we have checked knldiag/knldiagerr files and we are unable to find what is the root cause of the problem?

The root cause is the fact, that the parameter "UserTasks" is too small. Too many concurrent connections are there. If you increase that parameter you have to restart the Livecache.

Markus

[Former Member]

Thanks Markus for your insight.

MAxUsertaks was set to 500( ((/SAPAPO/OM13/LCA03 check are ok concerning all parameters)

Lc session should be equal to 3*max work process +4 (we have 160 wp)

but we tried to find the root cause of the bottleneck (the lack of task is just a side effect for me)

how do we possible find the origin of the process UKT? (which increase high CPU consumption )

(we have made livecache trace, but hard to read, sap trace st03n/stad..)

Best regards,

John

Edited by: DCWE COREADMIN on Feb 12, 2011 3:59 PM