work proccess restarts

[Former Member]

Hi,

I am heaving problem with restarting proccesses on ECC 5.0.

In SM50 some of the WP have been restarted 9 times in last 3 days, bellow is the error message in WP log:

(0) 0x40000000022fbb20 CTrcStack2 + 0x1b0 at dptstack.c:220 [dw.sapPRD_DVEBMGS01]

(1) 0x40000000022fb930 CTrcStack + 0x30 at dptstack.c:175 [dw.sapPRD_DVEBMGS01]

(2) 0x4000000003e2a950 rstsos__obj_stat + 0x280 at rstsow.c:333 [dw.sapPRD_DVEBMGS01]

(3) 0x4000000003e31c80 rstsro_repair_object + 0x2c0 at rstsow.c:2096 [dw.sapPRD_DVEBMGS01]

(4) 0x4000000003ca94c0 BtcTlgFcl + 0xa0 at btctlgcl.c:95 [dw.sapPRD_DVEBMGS01]

(5) 0x4000000003caa4d0 BtcTlgAp + 0x2b0 at btctlgap.c:54 [dw.sapPRD_DVEBMGS01]

(6) 0x4000000003c562d0 BtcLgAp + 0x50 at btclog.c:157 [dw.sapPRD_DVEBMGS01]

(7) 0x4000000003c61280 BtcCleanUp + 0x140 at btcjcntl.c:1248 [dw.sapPRD_DVEBMGS01]

(8) 0x400000000207a7f0 ThSUsrDel + 0x640 at thxxmode.c:4559 [dw.sapPRD_DVEBMGS01]

(9) 0x4000000002078340 ThIAMDel + 0xfe0 at thxxmode.c:2493 [dw.sapPRD_DVEBMGS01]

(10) 0x4000000002077280 ThAMDel + 0xa0 at thxxmode.c:2266 [dw.sapPRD_DVEBMGS01]

(11) 0x400000000205d200 ThCleanPrevUser + 0x770 at thxxhead.c:16752 [dw.sapPRD_DVEBMGS01]

(12) 0x40000000020157e0 ThCheckPrevUser + 0xd10 at thxxhead.c:13019 [dw.sapPRD_DVEBMGS01]

(13) 0x4000000001ff5f40 TskhLoop + 0x4150 at thxxhead.c:2915 [dw.sapPRD_DVEBMGS01]

(14) 0x4000000001feca40 tskhstart + 0x3a0 at thxxhead.c:1073 [dw.sapPRD_DVEBMGS01]

(15) 0x4000000001ebeb30 DpMain + 0x490 at dpxxdisp.c:1038 [dw.sapPRD_DVEBMGS01]

(16) 0x400000000166b1b0 nlsui_main + 0x30 [dw.sapPRD_DVEBMGS01]

(17) 0x400000000166b140 main + 0x70 at thxxanf.c:68 [dw.sapPRD_DVEBMGS01]

(18) 0xc000000000043880 main_opd_entry + 0x50 [/usr/lib/hpux64/dld.so]

L *** ERROR => BtcCleanUp: BtcLgAp-call failed (rc = 4) [btcjcntl.c 1251]

M

M

M Fri Dec 31 02:40:26 2010

M ***LOG Q0E=> SigIGenAction, signal ( 11) [sigux.c 903]

M in_ThErrHandle: 1

M ThSigHandler: signal (step 4, th_errno 11, action 4)

M

M Info for wp 26

M

M stat = 4

M reqtype = 4

M act_reqtype = 4

M rq_info = 8576

M tid = 26

M mode = 0

M len = 24

M rq_id = 23818

M rq_source = 1

M last_tid = 26

M last_mode = 0

M int_checked_resource(RFC) = 0

M ext_checked_resource(RFC) = 0

M int_checked_resource(HTTP) = 0

M ext_checked_resource(HTTP) = 0

M report = >SAPLTHFB <

M action = 11

M tab_name = >SHP_IDX_GDSI <

M

M Modeinfo for User T26/M0

M

M tm state = 2

M uid = 27

M term type = 0x2

M display = 0x0

M cpic_no = 0

M cpic_idx = -1

M usr = >PGI <

M terminal = > <

M client = >100<

M conversation_ID = > <

M appc_tm_conv_idx = -1

M its_plugin = NO

M allowCreateMode = YES

M blockSoftCanel = NO

M imode = 2

M mode state = 0x42

M mode clean_state = 1

M task_type = ZTTABTC

M th_errno = 11

M rollout_reason = 0

M last_rollout_level = 7

M async_receives = 0

M cpic_receive = 0

M em handle = 2

M roll state = 2

M abap state = 2

M em state = 3

Just to take this ot of the way , my current kernel release is 347. We have upgraded oracel from 9.2.0.8. to 10.2.0.4 , and applied the latest patches both CPU and interim.

Same reportsare taking much longer after upgrade.

Does anyone have any sugestion?

Thanks in advance.

D.

[former_member227281]

Hi,

Regarding work processes restarted many times, please check database parameters.

If possible increase the value of Processes and Sessions in Database parameters.

Regarding the performance that Same reportsare taking much longer after upgrade, it is very common everywhere.

There are many steps to be checked as follows

1. Database parameters recommendation as per the OSS note: 830576

2. Delete database statistics.

3. Then apply Database staistics for Dictionary , system and fixed objects as per the SAP oss note: 838725.

4. Then apply database statistics for all SAP objects as mentioned in Upgrade guide (see the Post processing section).

5. Then enter tx codes sm50 for each dialog instances and see which reports are taking long time. Note down the table name (See the Action column - Sequential read <table_name>.

6. Perform compute statistics for those tables mentioned in Sr. 5.

7. Check explan plan in current oracle sessions and see whether costs is high. If it reports full table scan then check index is present or not with where clause. Some indexes are not created before upgrade but the indexes are required after upgrade

If it showed index range scan and the costs is high, it could be incorrect index or inefficient index oraccess path is poor

If the access path is poor, then perform compute e statistics for those tables if the number of records are less than 50 lakhs otherwise use estimate statistics with 20% sampling size

8. Inform Abaper to fine tune the custom report. ( I had expereiced performance problems after upgrade).

Note: execute database statistics for some tables using dbms-stats-gather table for item no. 6 becasue Analyze table <table_name> compute is no longer supported in oracle 10g and higher.

Thanks and Regards

-apr

[Former Member]

Now that you upgraded to Oracle 10.2, did you have a chance to review the parameters for Oracle? [https://service.sap.com/sap/support/notes/1171650|Automated Oracle DB parameter check]

J. Haynes

[former_member189546]

Hello,

The issue might be the one described in

1458503 Blocking work processes: Many R0M syslog entries

regards,

John Feely

[Former Member]

Also one more thing from the log:

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

*

• LOCATION SAP-Server sap4_PRD_01 on host sap4 (wp 26)

• ERROR ThSigHandler: signal

*

• TIME Fri Dec 31 02:49:26 2010

• RELEASE 640

• COMPONENT Taskhandler

• VERSION 1

• RC 11

• MODULE thxxhead.c

• LINE 9993

• COUNTER 10

*