System is very Slow

[Former Member]

Dear all,

My Production server is running very slow.

I am sending ST03N log:

Dispatcher wait time(Wait Time)- 0.4 ms

Load+Gen Time- 4.5 ms

Roll-in Time- 0.6 ms

Roll Wait Time- 94.7 ms

Gui Time- 96.0 ms

Processing Time- 90.3 ms

CPU Time- 51.9 ms

Database Time- 486.8 ms

Response Time- 677.4 ms

Wait Time- 0.4 ms

Regards,

PC Patel

[Former Member]

This is very vast topic.

What is your OS and DB.

Is your stats are upto date?

also check in SM50 what action you see when user perform any transaction.

check for your disk operations (is there any IO wait)

also when you have identified system is slow? have you applied any patches or kernel upgrade?

[Former Member]

Dear Sunil,

OS is 2003 Server and DB is Oracle 10.0.2

State is up to date.

In SM50 8-15 work process are busy during slow, total Number of Dialog WP is 30

and all other WP like Enq,Upd,bgd,spo.....etc are free.

Regards,

PC Patel

[Former Member]

Hi,

15 dailog processes engaged at once it not normal. Please check if there are any reports are running at that particular time.

Else check which activity is making 15 work processes engaged. You have analyse that and then need to fix along with the owner of the activty.

Cheers...,

Raghu

[former_member227283]

Hi,

when you say, 15 dailog process are engaged, can you check what exactly going on the system. Depend upon which we can move forward for your system performance bottelnek.

Thanks

Anil

[Former Member]

hm! not look good from SM50, are these doing seq

(Sorry, some how this is saved with half info. see below)

Edited by: Sunil Bujade on Oct 20, 2010 9:30 AM

[Former Member]

Dear Anil,

Mostly System is going slow when 4-5 user run Tcode MB51 simultaneously.

When I check ST03N then I found DB time is very high(486 ms).

What should I check more.

Regards,

PC Patel

[Former Member]

hm! not look good from SM50, are these doing sequential read / direct read? is there any resource hungry job is active? also check your database is all looks OK? tablespace any alerts. Is backup still active?

[Former Member]

MB51 tcode doesn't have any mandatory fields, but then also system will show you a warning on task manager (SAP GUI). Please try to restrict your search selection by may be with material, plant, posting date etc, check with functional team about this and then monitor the system performance.

[former_member227283]

Hi,

Generally MB51 uses table MSEG, this is table which contain heavy data.

Can you check how much is a data present in table MSEG table. Is there any customize index created for this table. If no, then i will suggest to create index for table MSEG with field MANDT, LFBNR and LFPOS.

As you aware MB51 is nothing but to check the list of material document which basically make select on above fields.

Thanks

Anil

[Former Member]

Dear Sunil,

Sm50 doing sequential read, database is OK, No Tablespace error and No backup is running.

Daily Backup is run at 11.00 PM.

Regards,

PC Patel

[Former Member]

Dear Anil,

Two Index are already created with field

LFBNR

LFPOS

LFBJA

and

MANDT

EBELN

EBELP

Regards,

PC Patel

[former_member227283]

Hi,

Create new index with field MANDT, LFBNR & LFPOS.

Every field in index has different logic of reading the data.

Before creating the index, seat with the user who is facing thr problem and check what time it take to get the output of MB51.

Then create the index as above and then again run MB51 and now check what time it takes for getting the output.

Thanks

Anil

[Former Member]

Dear Anil,

Thanks for your suggestion.

I did this, but getting same problem.

How we identified which field we choose to create optimum Index.

Regards,

PC Patel

[JPReyes]

This is very basic stuff, I recommend you to do Workload Analysis course.

Regards

Juan

[Former Member]

PC,

Check if the statistics are up to date in tcode DB20 for tables flashing in SM50 and if not then create the statistics for them.

Check AVg(ms) in ST04->Details analysys->File system requests->AVg(ms) ,It should be less than 10ms for datafiles,if not then there's I/o problem and you need to relocate the datafiles to some other disk.

Also run sql trace in ST05 and check for the expensive sql statements.

Below link might help you in sql optimization.

https://websmp202.sap-ag.de/~sapidb/011000358700001413351999E#row_162

[Former Member]

Hi,

Have you taken a look for SAP Notes on the performance of MB51? I've come across a similar situation in the past with other transactions and I've sometimes found this is caused by inefficiencies in the transaction that can be corrected.

If not, try searching for "MB51 Performance" in the SAP Note search. You may find some that give advice on how to improve performance of the transaction. For example, https://service.sap.com/sap/support/notes/921164.

I usually try to take a look at these first of all and then carry out further analysis if there are no known issues.

Hope this helps.

Regards.

Steve.