SAP Installation using Zones

[Former Member]

Hello All,

We are in process of revamping the SAP architecture, we are having Sun T5120 Server, we have installed Sun 10 OS. We have also configured Solaris Zones on this server to install DEV and QAS instance on the same server. After successfull installation of SAP R/3 4.7 and Oracle 9i on both the zones, and also upgrade both the servers to ECC 6.0 with oracle 10g. when we start SAP on DEV and QAS simultaneously, everything hangs up due to resource consumption.

We have already assigned 2 core and 4 GB memory to each instance separately.

When we analyze the issue, we come to know that on primary zone (DEV) the oracle resources of QAS instance are also running be oradev user.

Can u guys please guide me that is there any additional settings required for SAP installation on Solaris zones or in virtualizaton concept ?

Please let me know for any other details.

Early response will highly appreciated.

Thanks,

Kshitiz Goyal,

BASIS Consultant

[Former Member · Accepted solution]

hi goyal,

how many WP on your DEV and QA ? Depend on what your setting is.

Don't assign too many WP on both instance otherwise it will stuck.

I think Markus is right on some point of view. But Markus, it's all depend on DEV and QA setting, right ?

ardhian

http://sapbasis.wordpress.com

http://ardhian.kioslinux.com

[Former Member · Accepted solution]

Hello Ardhian,

Thanks for your reply,

Work process on DEV instance are as follows:

7 DIA

2 UPD

5 BGD

1 SPO

1 UPD2

Work Process on QAS instance are as follows:

3 DIA

1 UPD

2 BGD

1 SPO

1 UPD2

Please let me know for any other settings required for DEV and QAS instance in case of virtualization.

Thanks,

Kshitiz Goyal

[markus_doehr2 · Accepted solution]

> how many WP on your DEV and QA ? Depend on what your setting is.

> Don't assign too many WP on both instance otherwise it will stuck.

The number of workprocesses is not that relevant. SAP make use of shared memory, all workprocesses attach to the "same" memory segments.

Markus

[markus_doehr2 · Accepted solution]

What is the output of the command

sappfpar pf=/usr/sap/<SID>/profile/<instanceprofile> check

as user adm on both instances?

What´s the value of the parameter "sga_max_size" on both databases?

Markus

[Former Member · Accepted solution]

Hi Markus,

Output on both the node is as follows :::

ON DEV (Global Zone)

eccdev:devadm 5% sappfpar pf=DEV_DVEBMGS00_eccdev check

===============================================================================

== Checking profile: DEV_DVEBMGS00_eccdev

===============================================================================

***WARNING: Unexpected parameter: auth/auth_number_in_userbuffer =5000

Shared memory disposition overview

================================================================

Shared memory pools

Key: 10 Pool

Size configured.....: 104000000 ( 99.2 MB)

Size min. estimated.: 99695426 ( 95.1 MB)

Advised Size........: 102000000 ( 97.3 MB)

Key: 40 Pool for database buffers

Size configured.....: 138000000 ( 131.6 MB)

Size min. estimated.: 132030328 ( 125.9 MB)

Advised Size........: 136000000 ( 129.7 MB)

Shared memories inside of pool 10

Key: 11 Size: 500000 ( 0.5 MB) Factory calender buffer

Key: 12 Size: 6000000 ( 5.7 MB) TemSe Char-Code convert Buf.

Key: 13 Size: 40500000 ( 38.6 MB) Alert Area

Key: 16 Size: 22400 ( 0.0 MB) Semaphore activity monitoring

Key: 17 Size: 2672386 ( 2.5 MB) Roll administration

Shared memories inside of pool 40

Key: 42 Size: 7056992 ( 6.7 MB) DB TTAB buffer

Key: 43 Size: 52980632 ( 50.5 MB) DB FTAB buffer

Key: 44 Size: 7914392 ( 7.5 MB) DB IREC buffer

Key: 45 Size: 4842392 ( 4.6 MB) DB short nametab buffer

Key: 46 Size: 20480 ( 0.0 MB) DB sync table

Key: 47 Size: 6145024 ( 5.9 MB) DB CUA buffer

Key: 48 Size: 300000 ( 0.3 MB) Number range buffer

Key: 49 Size: 2769392 ( 2.6 MB) Spool admin (SpoolWP+DiaWP)

Shared memories outside of pools

Key: 1 Size: 2500 ( 0.0 MB) System administration

Key: 2 Size: 5601560 ( 5.3 MB) Disp. administration tables

Key: 3 Size: 13714400 ( 13.1 MB) Disp. communication areas

Key: 4 Size: 512248 ( 0.5 MB) statistic area

Key: 6 Size: 319488000 ( 304.7 MB) ABAP program buffer

Key: 7 Size: 14838 ( 0.0 MB) Update task administration

Key: 8 Size: 67108964 ( 64.0 MB) Paging buffer

Key: 9 Size: 134217828 ( 128.0 MB) Roll buffer

Key: 14 Size: 4400128 ( 4.2 MB) Presentation buffer

Key: 18 Size: 917604 ( 0.9 MB) Paging adminitration

Key: 19 Size: 30000128 ( 28.6 MB) Table-buffer

Key: 30 Size: 18432 ( 0.0 MB) Taskhandler runtime admin.

Key: 31 Size: 4806000 ( 4.6 MB) Dispatcher request queue

Key: 33 Size: 10240000 ( 9.8 MB) Table buffer, part.buffering

Key: 34 Size: 4194304 ( 4.0 MB) Enqueue table

Key: 41 Size: 25010000 ( 23.9 MB) DB statistics buffer

Key: 51 Size: 3200000 ( 3.1 MB) Extended memory admin.

Key: 52 Size: 40000 ( 0.0 MB) Message Server buffer

Key: 54 Size: 12492800 ( 11.9 MB) Export/Import buffer

Key: 55 Size: 8192 ( 0.0 MB) Spool local printer+joblist

Key: 57 Size: 1048576 ( 1.0 MB) Profilparameter in shared mem

Key: 58 Size: 4096 ( 0.0 MB) Enqueue ID for reset

Key: 62 Size: 85983232 ( 82.0 MB) Memory pipes

Key: 63 Size: 409600 ( 0.4 MB) ICMAN shared memory

Key: 64 Size: 4202496 ( 4.0 MB) Online Text Repository Buf.

Key: 65 Size: 4202496 ( 4.0 MB) Export/Import Shared Memory

Key: 1002 Size: 400000 ( 0.4 MB) Performance monitoring V01.0

Key: 58900100 Size: 4096 ( 0.0 MB) SCSA area

Nr of operating system shared memory segments: 30

Shared memory resource requirements estimated

================================================================

Total Nr of shared segments required.....: 30

System-imposed number of shared memories.: 223

Shared memory segment size required min..: 319488000 ( 304.7 MB)

System-imposed maximum segment size......: 15367929856 (14656.0 MB)

Swap space requirements estimated

================================================

Shared memory....................: 927.2 MB

..in pool 10 95.1 MB, 95% used

..in pool 40 125.9 MB, 95% used

..not in pool: 698.3 MB

Processes........................: 149.0 MB

Extended Memory .................: 2040.0 MB

-

---

Total, minimum requirement.......: 3116.2 MB

Process local heaps, worst case..: 1908.0 MB

Total, worst case requirement....: 5024.2 MB

Errors detected..................: 0

Warnings detected................: 1

eccdev:devadm 6%

*****************************SGA ON DEV DB********************************

Total System Global Area 4898947072 bytes

Fixed Size 2031168 bytes

Variable Size 2599322048 bytes

Database Buffers 2282848256 bytes

Redo Buffers 14745600 bytes

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

ON QAS (Local Zone)

eccqas:qasadm 1% sappfpar pf=/sapmnt/QAS/profile/QAS_DVEBMGS00_eccqas check

==========================================================================

== Checking profile: /sapmnt/QAS/profile/QAS_DVEBMGS00_eccqas

==========================================================================

***ERROR: Size of shared memory pool 10 too small

================================================================

SOLUTIONS: (1) Locate shared memory segments outside of pool 10

with parameters like: ipc/shm_psize_<key> =0

SOLUTION: Increase size of shared memory pool 10

with parameter: ipc/shm_psize_10 =108000000

***ERROR: Size of shared memory pool 40 too small

================================================================

SOLUTIONS: (1) Locate shared memory segments outside of pool 40

with parameters like: ipc/shm_psize_<key> =0

SOLUTION: Increase size of shared memory pool 40

with parameter: ipc/shm_psize_40 =112000000

Shared memory disposition overview

================================================================

Shared memory pools

Key: 10 Pool

Size configured.....: 68000000 ( 64.8 MB)

Size min. estimated.: 104095554 ( 99.3 MB)

Advised Size........: 108000000 ( 103.0 MB)

Key: 40 Pool for database buffers

Size configured.....: 68000000 ( 64.8 MB)

Size min. estimated.: 108776088 ( 103.7 MB)

Advised Size........: 112000000 ( 106.8 MB)

Shared memories inside of pool 10

Key: 11 Size: 500000 ( 0.5 MB) Factory calender buffer

Key: 12 Size: 6000000 ( 5.7 MB) TemSe Char-Code convert Buf.

Key: 13 Size: 40500000 ( 38.6 MB) Alert Area

Key: 14 Size: 4400000 ( 4.2 MB) Presentation buffer

Key: 16 Size: 22400 ( 0.0 MB) Semaphore activity monitorin

Key: 17 Size: 2672386 ( 2.5 MB) Roll administration

Shared memories inside of pool 40

Key: 42 Size: 7232992 ( 6.9 MB) DB TTAB buffer

Key: 43 Size: 32534392 ( 31.0 MB) DB FTAB buffer

Key: 44 Size: 7958392 ( 7.6 MB) DB IREC buffer

Key: 45 Size: 4886392 ( 4.7 MB) DB short nametab buffer

Key: 46 Size: 20480 ( 0.0 MB) DB sync table

Key: 47 Size: 3073024 ( 2.9 MB) DB CUA buffer

Key: 48 Size: 300000 ( 0.3 MB) Number range buffer

Key: 49 Size: 2769392 ( 2.6 MB) Spool admin (SpoolWP+DiaWP)

Shared memories outside of pools

Key: 1 Size: 2500 ( 0.0 MB) System administration

Key: 2 Size: 3381324 ( 3.2 MB) Disp. administration tables

Key: 3 Size: 13714400 ( 13.1 MB) Disp. communication areas

Key: 4 Size: 506848 ( 0.5 MB) statistic area

Key: 6 Size: 159744000 ( 152.3 MB) ABAP program buffer

Key: 7 Size: 14838 ( 0.0 MB) Update task administration

Key: 8 Size: 67108964 ( 64.0 MB) Paging buffer

Key: 9 Size: 134217828 ( 128.0 MB) Roll buffer

Key: 18 Size: 917604 ( 0.9 MB) Paging adminitration

Key: 19 Size: 30000000 ( 28.6 MB) Table-buffer

Key: 30 Size: 9216 ( 0.0 MB) Taskhandler runtime admin.

Key: 31 Size: 4806000 ( 4.6 MB) Dispatcher request queue

Key: 33 Size: 10240000 ( 9.8 MB) Table buffer, part.buffering

Key: 34 Size: 4194304 ( 4.0 MB) Enqueue table

Key: 41 Size: 25010000 ( 23.9 MB) DB statistics buffer

Key: 51 Size: 3200000 ( 3.1 MB) Extended memory admin.

Key: 52 Size: 40000 ( 0.0 MB) Message Server buffer

Key: 54 Size: 4202496 ( 4.0 MB) Export/Import buffer

Key: 55 Size: 8192 ( 0.0 MB) Spool local printer+joblist

Key: 57 Size: 1048576 ( 1.0 MB) Profilparameter in shared me

Key: 58 Size: 4096 ( 0.0 MB) Enqueue ID for reset

Key: 62 Size: 85983232 ( 82.0 MB) Memory pipes

Key: 63 Size: 409600 ( 0.4 MB) ICMAN shared memory

Key: 64 Size: 4202496 ( 4.0 MB) Online Text Repository Buf.

Key: 65 Size: 4202496 ( 4.0 MB) Export/Import Shared Memory

Key: 1002 Size: 400000 ( 0.4 MB) Performance monitoring V01.0

Key: 58900100 Size: 4096 ( 0.0 MB) SCSA area

Nr of operating system shared memory segments: 29

Shared memory resource requirements estimated

================================================================

Total Nr of shared segments required.....: 29

System-imposed number of shared memories.: 256

Shared memory segment size required min..: 159744000 ( 152.3 MB)

System-imposed maximum segment size......: 15435038720 (14720.0 MB)

Swap space requirements estimated

================================================

Shared memory....................: 743.5 MB

..in pool 10 64.8 MB, 153% used !!

..in pool 40 64.8 MB, 159% used !!

..not in pool: 531.7 MB

Processes........................: 75.1 MB

Extended Memory .................: 4092.0 MB

-

---

Total, minimum requirement.......: 4910.6 MB

Process local heaps, worst case..: 1907.3 MB

Total, worst case requirement....: 6818.0 MB

Errors detected..................: 2

Warnings detected................: 0

eccqas:qasadm 2%

*******************SGA on QAS DB************************

Total System Global Area 4784723408 bytes

Fixed Size 734672 bytes

Variable Size 2499805184 bytes

Database Buffers 2282848256 bytes

Redo Buffers 1335296 bytes

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

Thanks,

Kshitiz Goyal

[markus_doehr2 · Accepted solution]

Your memory is already overcomitted just by the database memory:

> DEV

> Total System Global Area 4898947072 bytes

> QAS

> Total System Global Area 4784723408 bytes

You said you assigned 4 GB per instance. This means as soon as the database starts, the system is using swap already (900 MB/800 MB). If you then start the ABAP instance on top, it will allocate additional 3 GB memory per instance (according to your profile).

Since Oracle uses DISM and this memory is not paged out, the whole ABAP stack will JUST work with paging - which makes the system very very slow. Every click will lead to an I/O on the system.

The problem is not caused by virtualization and/or zones but by your configuration.

Markus

[Former Member · Accepted solution]

Markus,

Thanks for giving this much information, your suggestions will really help us a lot.

Can u please guide me about what configuration should i adopt, or please give me some notes to proceed futher.

Thanks,

Kshitiz

[markus_doehr2 · Accepted solution]

If you have more memory I would increase it per zone to 8 GB. A 6.20 based 4.7 system could run pretty well with 4 GB, running an ERP 6.0 system with that few memory is almost impossible and no fun for people working with it.

If you don´t have the memory decrease sga_max_size to 2 GB and check your parameters. Do not edit the parameter files manually on OS level but use RZ10. This transaction will automatically caluclate the shared memory sizes necessary. However, with 4 GB you won´t become happy with speed, the system will still be slow.

Make sure the sga_max_size + initial extended memory together don´t exceed your physical available memory per instance.

Markus

[markus_doehr2]

> We have already assigned 2 core and 4 GB memory to each instance separately.

4 GB for a full zone including instance plus upgrade? That WAY too less memory.

By default the databases uses half of the available memory, that is 2 GB, now your 4.7 instance checks the available memory (2 GB + swap) and allocates 25 % of it. On Solaris DISM is used means, the database memory is not swapped out. Now you start the upgrade and a shadow instance which does the same (25 % of memory + swap).

If your systems are Unicode the situation is even worse.

Assign more memory to the instances, otherwise the whole system will keep swapping heavily.

Markus

[Former Member]

Hi Markus,

Thanks for your reply,

I would like to mention some more points here,

1. we are using non-unicode system.

2. Upgrade from R/3 4.7 to ECC 6.0 SR3 finishes successfully on both the instance separately.

3. When we down the QAS instance, and run only DEV instance with 2 CPU Core and 4 GB memory, it works perfectly fine.

4. But when we start the QAS zone also, and view the processes from primary Zone (DEV) through

ps -ef | grep command, oracle processes of QAS zone are also run using oradev user.

oradev 23204 1 0 06:00:52 ? 0:01 ora_mman_DEV

oradev 23210 1 0 06:00:53 ? 0:02 ora_dbw2_DEV

oradev 23212 1 0 06:00:53 ? 0:02 ora_dbw3_DEV

oradev 23220 1 0 06:00:53 ? 0:00 ora_reco_DEV

oradev 23218 1 0 06:00:53 ? 0:13 ora_smon_DEV

oradev 23208 1 0 06:00:52 ? 0:02 ora_dbw1_DEV

oradev 23202 1 0 06:00:52 ? 0:01 ora_psp0_DEV

oradev 23216 1 0 06:00:53 ? 0:10 ora_ckpt_DEV

oradev 23222 1 0 06:00:53 ? 0:03 ora_mmon_DEV

oradev 23200 1 0 06:00:52 ? 0:15 ora_pmon_DEV

oradev 23256 1 0 06:00:59 ? 0:04 ora_arc0_DEV

oradev 23268 1 0 06:01:02 ? 0:00 ora_qmnc_DEV

oradev 23214 1 0 06:00:53 ? 0:05 ora_lgwr_DEV

oradev 23319 1 0 06:01:54 ? 0:00 ora_q001_DEV

oradev 23306 1 0 06:01:12 ? 0:01 ora_q000_DEV

oradev 23258 1 0 06:00:59 ? 0:00 ora_arc1_DEV

oradev 23224 1 0 06:00:53 ? 0:29 ora_mmnl_DEV

oradev 23206 1 0 06:00:52 ? 0:02 ora_dbw0_DEV

oradev 23207 1 0 06:00:52 ? 0:01 ora_mman_QAS

oradev 24210 1 0 06:00:53 ? 0:02 ora_dbw2_QAS

oradev 24212 1 0 06:00:53 ? 0:02 ora_dbw3_QAS

oradev 24220 1 0 06:00:53 ? 0:00 ora_reco_QAS

oradev 24218 1 0 06:00:53 ? 0:13 ora_smon_QAS

oradev 24209 1 0 06:00:52 ? 0:02 ora_dbw1_QAS

oradev 24201 1 0 06:00:52 ? 0:01 ora_psp0_QAS

oradev 24218 1 0 06:00:53 ? 0:10 ora_ckpt_QAS

oradev 24224 1 0 06:00:53 ? 0:03 ora_mmon_QAS

Please suggest from your experience, why it is happening ?

Thanks a lot,

Kshitiz Goyal