on 07-17-2013 10:08 AM
Hi everybody.
We have notice poor performance in our system. At the same time job SAP_COLLECTOR_FOR_PERFMONITOR ends with errors 2 times every day, at 04:00 and at 20:00 everyday with dump TSV_TNEW_PAGE_ALLOC_FAILED.
Looking at ST02 I can see that all the Hit Ratio buffers are over 99% but swaps are mustly in red.
Can anybody tell me if this is normal behaviour or if I need to increase some parameters?.
Attached I send you an ST02 screenshot.
Thanks in advance for your help and best regards.
Hi Deepak,
I would increase the InitialRecord and Export/Import as the hit ratio is too low for these. Your hit ratios should be 95% or better.
Also check note #785038, this note is a little outdated but it may give you an idea of what you need to do like download a support pack.
And note #1110822
Cheers,
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Francisco,
1) at 04:00 and at 20:00 check are there any long running background jobs ?
2) what is the SAP release and what is the server configuration ?
Regards,
Deepak Kori
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Deepak.
At the time the job is running there are no other long running background jobs (max 47 seconds of execution).
My SAP release es ECC 6.0, kernel 7.20 patch 401, SAP_ABA, SAP_BASIS, etc release 7.01, level 10. The server configuration is an iSeries OS/400 release 6.1 in an LPAR with 3 processes units and 44,5 GB of memory asigned. The ASP has 1557 GB and is 76,57% in use.
The SAP instance has 41 work processes with two operations configuration:
-DAY: From 7am to 8pm: 20 dialog, 10 batch, 2 batchA, 4 spool, 4 update, 2 update2, 1 enqueue
-NIGHT: From 20pm to 7am: 5 dialog, 25 batch, 2 batchA, 4 spool, 4 update, 2 update2, 1 enqueue
Thanks for your help.
Hi Francisco,
Check note 1110822.
Also check whether any load is there in the system during that time when the job is getting failed.
Primarily it is due to exceeding the size of memory , please also let us know the profile parameter settings for memory parameters in your system.
Thanks and Regards,
Kunal Gahlot
Hi Francisco,
Looking at ST02 results , there is no enough free space for following buffers
table definition
program
screen
calendar
Export/Import
You can double click on the respective entries to identify SAP profile parameters associated with them,
Try to double the value and save the profile file
check the configuration of shared memory using sappfpar -check pf=<path of instance profile>
Ensure there are no errors.
Take SAP restart and monitor the buffer settings in ST02
Hope this helps.
Regards,
Deepak Kori
Hi Deepak.
This is what I'm going to change:
Parameter Old Value Increment New Value
rsdb/ntab/entrycount 39000 15600 54600
rsdb/ntab/ftabsize 80000 32000 112000
rsdb/ntab/irbdsize 9500 3800 13300
abap/buffersize 960000 384000 1344000
zcsa/presentation_buffer_area 20000000 8000000 28000000
sap/bufdir_entries 10000 4000 14000
zcsa/table_buffer_area 100000000 40000000 140000000
zcsa/db_max_buftab 10000 4000 14000
rtbb/buffer_length 60000 24000 84000
rtbb/max_tables 500 200 700
rsdb/obj/buffersize 128000 51200 179200
rsdb/obj/max_objects 32000 12800 44800
rsdb/obj/large_object_size 8192 3276,8 11468,8
ztta/roll_area 3000000 1200000 4200000
em/initial_size_MB 8192 3276,8 11264
User | Count |
---|---|
88 | |
23 | |
11 | |
9 | |
8 | |
5 | |
5 | |
5 | |
5 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.