Solved: Degrade performance in ECC 6.0

Hi,

in a system upgraded previus month, form few days we have bad performance. It's a Solaris 10 SPARC box with Oracle 10.2.0.4. This is a ECC 6.0 Abap only system with Kernel patch 180.

Yesterday eveneing we try with a "brconnect -c -u / -f stats -t all -f collect -p 4" to have opt. statistic but without success.

Until previus week we have 700-800 ms as dialog responde time, from saturday we have 1500-1600 ms as dialog responce time. And several btc job run for a long time.

Have you got any advice ?

Regards.

You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.

Comment

Accepted Solutions (1)

- did you install all the interim patches from Note 1137346 - Oracle Database 10g: Patches for Release 10.2.0.4?

- are all parameters set according to Note 830576 - Parameter recommendations for Oracle 10g (especially the FIXCONTROL and event parameters?

Markus

You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.

Comment

Answers (1)

maybe just a few SQL requests are causing this.

check in SM50 for long running SQL requests. when you find the culprit, analyze explain plan for this SQL request.

are users complaining?

You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.

Comment

Hi,

I find a very large Kernel CPU usage, User CPU is quite normal...

Does /var/adm/messages tell you something? What about full devices? What is the situation of your I/O devices?

André

100 should be ora<sid>

101 should be <sid>adm

32 Gb phisical ram.... 100 id 21% ? 101 19% ? ... other mem ?

regards....

NPROC USERNAME SWAP RSS MEMORY TIME CPU

83 101 24G 6232M 19% 5:42:02 14%

93 100 6966M 6895M 21% 0:40:21 5.3%

160 root 282M 282M 0.9% 0:20:49 4.3%

20 daemon 21M 29M 0.1% 0:04:40 0.4%

3 noaccess 426M 370M 1.1% 0:02:13 0.0%

1 109 23M 26M 0.1% 0:00:01 0.0%

7 103 30M 36M 0.1% 0:00:30 0.0%

1 lp 1040K 1584K 0.0% 0:00:00 0.0%

1 102 4064K 25M 0.1% 0:00:01 0.0%

3 smmsp 4128K 11M 0.0% 0:00:00 0.0%

Sun suggest us to put into /etc/system of global zone this param...

set pg_contig_disable=1

We try to apply in week-end.

Regards.

Hi,

>Does /var/adm/messages tell you something? What about full devices? What is the situation of your I/O >devices?

no error or particular other.

Regards.

Ganimede Dignan

Hi,

noone use pg_contig_disable=1 on a Solaris 10 box ?

Regards.

>Sun suggest us to put into /etc/system of global zone this param...

>set pg_contig_disable=1

>We try to apply in week-end.

>Regards.

> set pg_contig_disable=1

Oracle tries to allocate huge blocks continually without fragmentation. If you set that parameter this feature is switched off.

Which filesystem is underlying? UFS? VxFS? ZFS?

Markus

Hi,

>Which filesystem is underlying? UFS? VxFS? ZFS?

Is it important ?

All datafiles, ora binary, sap binary, ect. ara on UFS filesystem. BRBACKUP destination ara on disk with ZFS filesystem.

Regards.

> Is it important ?

Of course the underlying filesystem and its configuration is important

> All datafiles, ora binary, sap binary, ect. ara on UFS filesystem. BRBACKUP destination ara on disk with ZFS filesystem.

Check the following notes:

Note 892631 - Solaris: mount options for best performance

Note 999524 - Oracle 10.2 platform support for Async IO

If you see a very high kernel CPU time you can try to find out which syscall is taking so long by doing a

pstack <PID>

Please also post some output of

vmstat 1

(omit the first output line)

Markus

Hi Markus,

all filesystem are mounted by zone (zonfigured into zonecfg):

<?xml version="1.0" encoding="UTF-8"?>

<!DOCTYPE zone PUBLIC "-//Sun Microsystems Inc//DTD Zones//EN" "file:///usr/share/lib/xml/dtd/zonecfg.dtd.1">

<!--

DO NOT EDIT THIS FILE. Use zonecfg(1M) instead.

-->

<filesystem special="/dev/dsk/c5t600A0B800032DB3E0000069B47600BF2d0s0" raw="/dev/rdsk/c5t600A0B800032DB3E0000069B47600BF2d0s0" directory="/oracle" type="ufs

"/>

<filesystem special="/dev/dsk/c5t600A0B800032DB060000059247544ADFd0s0" raw="/dev/rdsk/c5t600A0B800032DB060000059247544ADFd0s0" directory="/oracle/<SID>/saparch

h" type="ufs"/>

<filesystem special="/dev/dsk/c5t600A0B800032DB060000057B475441F9d0s0" raw="/dev/rdsk/c5t600A0B800032DB060000057B475441F9d0s0" directory="/oracle/<SID>/sapdata1

a1" type="ufs"/>

<filesystem special="/dev/dsk/c5t600A0B800032DB0600000595475450A9d0s0" raw="/dev/rdsk/c5t600A0B800032DB0600000595475450A9d0s0" directory="/oracle/<SID>/sapdata2

a10" type="ufs"/>

<filesystem special="/dev/dsk/c5t600A0B800032DB3E000005A847544680d0s0" raw="/dev/rdsk/c5t600A0B800032DB3E000005A847544680d0s0" directory="/oracle/<SID>/sapdata3

a11" type="ufs"/>

<filesystem special="/dev/dsk/c5t600A0B800032DB060000058A475446ABd0s0" raw="/dev/rdsk/c5t600A0B800032DB060000058A475446ABd0s0" directory="/oracle/<SID>/sapdata4

a12" type="ufs"/>

<filesystem special="/dev/dsk/c5t600A0B800032DB060000058C475446E7d0s0" raw="/dev/rdsk/c5t600A0B800032DB060000058C475446E7d0s0" directory="/oracle/<SID>/sapdata5

a13" type="ufs"/>

<filesystem special="/dev/dsk/c5t600A0B800032DB3E000005974754428Ed0s0" raw="/dev/rdsk/c5t600A0B800032DB3E000005974754428Ed0s0" directory="/oracle/<SID>/sapdata6

a2" type="ufs"/>

<filesystem special="/dev/dsk/c5t600A0B800032DB060000057D475442E1d0s0" raw="/dev/rdsk/c5t600A0B800032DB060000057D475442E1d0s0" directory="/oracle/<SID>/sapdata7

a3" type="ufs"/>

...

</filesystem>

....

This is vmstat output during problem

37 0 0 51868192 820712 266 3231 876 42 42 0 0 3 0 0 2 6510 29340 11284 18 82 0

35 0 0 51989520 896568 81 1988 17 43 43 0 0 1 0 0 4 8063 59828 14016 27 73 0

19 0 0 52012632 891504 388 6772 387 61 61 0 0 12 0 0 2 12986 221278 21347 47 53 0

9 0 0 51973352 854936 99 2014 67 20 20 0 0 2 0 1 1 15112 148334 25074 47 51 1

2 0 0 51897000 818816 13 914 10 51 51 0 0 1 0 0 3 20053 222209 29337 46 39 15

0 0 0 51934800 856032 67 1792 13 68 68 0 0 1 0 6 2 16630 302977 24303 30 30 41

0 0 0 51894216 840104 10 343 3 39 39 0 0 1 0 2 2 20151 292123 28203 26 14 60

0 0 0 51926032 868944 11 262 1 26 26 0 0 1 0 2 1 20961 274965 29301 24 20 57

0 0 0 51923304 858680 12 342 2 3 3 0 0 1 0 2 0 19689 265411 27941 16 33 51

...

0 0 0 51893176 833888 45 765 2 18 17 0 0 3 0 0 2 21429 61307 31652 21 38 41

5 0 0 51941936 861136 29 604 111 39 39 0 0 3 0 2 2 17090 58435 26988 25 56 19

6 0 0 51918216 885328 237 2020 344 18 18 0 0 2 0 1 2 16559 61083 26517 26 62 12

6 0 0 51901096 938816 59 771 133 25 25 0 0 2 0 3 3 18547 78140 30226 37 57 7

30 0 0 51894696 854800 12 706 40 64 63 0 0 1 0 2 4 4612 53445 8850 13 87 0

23 0 0 51988024 872344 12 965 41 38 38 0 0 1 0 2 2 7962 43598 14882 23 77 0

36 0 0 51994008 835552 5 671 12 7 6 0 0 1 0 0 2 5348 27493 10552 18 82 0

39 0 0 52008184 817168 5 702 210 5 5 0 0 1 0 2 1 5008 23290 9392 17 83 0

35 0 0 52052504 847776 33 938 10 52 51 0 0 1 0 0 2 6882 35230 12600 21 79 0

...

30 0 0 53391440 1337208 96 1162 136 124 124 0 0 1 0 2 0 13915 63083 23410 30 66 4

0 0 0 48757896 4151008 1923 1093 206 25 43 0 1909 1 0 2 1 27756 96296 41122 26 12 61

0 0 0 58978632 4865432 323 2599 430 100 99 0 0 8 0 4 1 21449 83337 31743 29 12 59

0 0 0 62863288 6068048 2296 846 180 8 6 0 0 1 0 1 0 17788 59524 25796 22 10 68

0 0 0 22012552 7000688 34 247 1 20 20 0 0 1 0 2 0 19917 68517 31344 13 8 78

Regards.

>Check the following notes:

>Note 892631 - Solaris: mount options for best performance

>Note 999524 - Oracle 10.2 platform support for Async IO

filesystemio_options=setall already setted.

Regards.

Your system times are REALLY very high, this is really not normal...

Do you see anything unusual when you execute

prstat -Z 1

on the global zone (maybe process fsflush)? Is it refreshing nicely every second or do you see interruptions?

I have seen such high system time only when lots for fork()s are happening, we had that once when a script was re-executed due to a wrong programmed loop. I have also seen it when huge memory portions are allocated by a program and free´ed from the filesystem cache.

Markus

Solved, with Sun parma suggested.

regards.

Hi,

are there any other significat patch to install for a ECC 6.0 on Oracle 10.2.0.4 ?

Regards.

Ganimede Dignan.

Installed Top-level Products (2):

Oracle Database 10g 10.2.0.1.0

Oracle Database 10g Release 2 Patch Set 3 10.2.0.4.0

There are 2 products installed in this Oracle Home.

Interim patches (27) :

Patch 7228971 : applied on Fri Nov 14 16:36:18 CET 2008

Created on 8 Jul 2008, 11:44:24 hrs PST8PDT

Bugs fixed:

6440977, 6120483, 5714944, 6626018, 6670551, 6329318, 7149751, 7228971

6988848, 7188932, 6221403, 6972291

Patch 7149863 : applied on Fri Nov 14 16:35:42 CET 2008

Created on 9 Jun 2008, 02:59:31 hrs PST8PDT

Bugs fixed:

7149863, 7022905, 5623467

Patch 7046187 : applied on Fri Nov 14 16:35:21 CET 2008

Created on 21 May 2008, 10:46:23 hrs US/Pacific

Bugs fixed:

5890312, 6760697, 6962622, 5970301, 7046187, 5934363, 5704108

Patch 7027551 : applied on Fri Nov 14 16:34:56 CET 2008

Created on 4 Jun 2008, 10:50:09 hrs US/Pacific

Bugs fixed:

7027551

Patch 6972843 : applied on Fri Nov 14 16:34:38 CET 2008

Created on 21 May 2008, 11:10:01 hrs US/Pacific

Bugs fixed:

6972843

Patch 6954829 : applied on Fri Nov 14 16:34:20 CET 2008

Created on 21 May 2008, 03:53:49 hrs US/Pacific

Bugs fixed:

6954829

Patch 6944036 : applied on Fri Nov 14 16:33:59 CET 2008

Created on 5 Jun 2008, 02:16:55 hrs PST8PDT

Bugs fixed:

6944036

Patch 6934892 : applied on Fri Nov 14 16:33:50 CET 2008

Created on 24 Jun 2008, 23:47:13 hrs PST8PDT

Bugs fixed:

6934892

Patch 6917874 : applied on Fri Nov 14 16:33:39 CET 2008

Created on 13 May 2008, 08:21:36 hrs US/Pacific

Bugs fixed:

6917874

Patch 6826661 : applied on Fri Nov 14 16:33:16 CET 2008

Created on 5 Jun 2008, 02:21:26 hrs PST8PDT

Bugs fixed:

6826661

Patch 6768114 : applied on Fri Nov 14 16:33:06 CET 2008

Created on 22 May 2008, 12:01:38 hrs US/Pacific

Bugs fixed:

6768114

Patch 6705635 : applied on Fri Nov 14 16:32:53 CET 2008

Created on 14 May 2008, 22:50:58 hrs PST8PDT

Bugs fixed:

6705635

Patch 6679303 : applied on Fri Nov 14 16:32:45 CET 2008

Created on 8 Jul 2008, 22:19:26 hrs PST8PDT

Bugs fixed:

6679303

Patch 6645719 : applied on Fri Nov 14 16:32:32 CET 2008

Created on 22 May 2008, 12:12:26 hrs US/Pacific

Bugs fixed:

6645719

Patch 6506617 : applied on Fri Nov 14 16:32:19 CET 2008

Created on 11 Jun 2008, 10:28:30 hrs US/Pacific

Bugs fixed:

6506617

Patch 6396242 : applied on Fri Nov 14 16:31:57 CET 2008

Created on 17 Mar 2008, 23:02:47 hrs US/Pacific

Bugs fixed:

6396242

Patch 6324944 : applied on Fri Nov 14 16:31:40 CET 2008

Created on 21 May 2008, 12:42:06 hrs US/Pacific

Bugs fixed:

6324944

Patch 6151380 : applied on Fri Nov 14 16:31:19 CET 2008

Created on 21 May 2008, 07:02:06 hrs US/Pacific

Bugs fixed:

6151380

Patch 6084232 : applied on Fri Nov 14 16:30:48 CET 2008

Created on 22 May 2008, 12:36:51 hrs US/Pacific

Bugs fixed:

6084232

Patch 6051177 : applied on Fri Nov 14 16:30:30 CET 2008

Created on 21 May 2008, 09:42:56 hrs US/Pacific

Bugs fixed:

6051177

Patch 5935935 : applied on Fri Nov 14 16:30:10 CET 2008

Created on 21 May 2008, 04:25:22 hrs US/Pacific

Bugs fixed:

5935935

Patch 5923486 : applied on Fri Nov 14 16:29:50 CET 2008

Created on 21 May 2008, 05:55:44 hrs US/Pacific

Bugs fixed:

5923486

Patch 5895190 : applied on Fri Nov 14 16:29:28 CET 2008

Created on 21 May 2008, 05:38:52 hrs US/Pacific

Bugs fixed:

5895190

Patch 5756769 : applied on Fri Nov 14 16:29:07 CET 2008

Created on 22 May 2008, 00:35:33 hrs US/Pacific

Bugs fixed:

5756769

Patch 5747462 : applied on Fri Nov 14 16:28:41 CET 2008

Created on 22 May 2008, 00:17:34 hrs US/Pacific

Bugs fixed:

5747462

Patch 5386204 : applied on Fri Nov 14 16:28:14 CET 2008

Created on 19 May 2008, 12:37:10 hrs US/Pacific

Bugs fixed:

5386204

Patch 5251842 : applied on Fri Nov 14 16:27:51 CET 2008

Created on 2 Jul 2008, 11:27:18 hrs US/Pacific

Bugs fixed:

5251842

Ask a Question

Degrade performance in ECC 6.0

Accepted Solutions (1)

Accepted Solutions (1)

Answers (1)

Answers (1)

Re: ASEGURAMIENTO DE TABLAS

Re: Getting error while Creating role in BRM

Re: How to enable the add, delete,edit system on s...

Re: CJS-30252 Running priviledged function stop of...

Successfactors login issue.