on 11-03-2007 11:41 AM
Hi Expert!
We have above 250 users in our Production Server and always we have complain of slow performance, our most critial transactions are VF01, VL01, VL02N and VA01.
I am writing the Policy for the Application Server to shutdow / restart once time a month / 15 days. I want to know pros / cons before implement the policy to my company.
1) Restart SAP Instances
2) Restarting SAP Server.
What is your expereince and your recommendation to start the SAP instances or Restarting the SAP Server. ?
Does SAP recommend the above task ?
Points will be awarded.
Regards
Hi,
1) Restarting SAP instance will empty it's buffers, thus slowing (till buffers get filled again) your instance.
2) Restarting SAP serwer will empty DB buffers (SAP too), thus slowing down entire system.
I do not know about SAP recommendation, but I recommend restarting servers only when absolutelly necessary (hardware problems/replacements, significant changes to instance profiles). I have SAP server (central instance) with uptime of 639 days and instance uptime of more than 100 days and it works just fine.
Try rather tuning your profile parameters (or adding hardware components).
Regs,
=FS=
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi
We have slow performance issues therefore we are deciding the Restart the SAP PRD server,
Please go through the my message down and give us the recommendations.
HI
We are using ECC 6 with Win2003 SP1 64 bit on Itanium Server with SQL2005 SP1 our highest patch level is 700 11 and kernel is 700 SP level 100.
We have 8GB RAM with 20GB page size and 3.5 GB is free on the local drive.
Our main business is Retail and Distribution; Our 25 users are working on SD documents out of 220 concurrent online users.
The 25 users are professional users which are working on the VA01 / VF01 / VL01n / VL02n. These transaction codes are utilizing the lot of resources of CPU, Memory and DB, and they are complaining the slow response of the server while others users are execute any processing for the reports or query.
We have set the Work Process as follows;
DIA 14
UPD 04
BGD 04
SPO 01
UP2 01
Our parameter values as follows
abap/buffersize 400000 kB Size of program buffer |
abap/pxa shared Program buffer mode |
rsdb/cua/buffersize 10000 kB Size of CUA buffer |
zcsa/presentation_buffer_area 20000768 Byte Size of screen buffer |
sap/bufdir_entries 10000 Max. number of buffered screens |
-
|
zcsa/table_buffer_area 100000768 Byte Size of generic key table buffer |
zcsa/db_max_buftab 10000 Max. number of buffered objects |
rtbb/buffer_length 60000 kB Size of single record table buffer |
rtbb/max_tables 500 Max. number of buffered tables |
rsdb/obj/buffersize 40000 kB Size of export/import buffer |
rsdb/obj/max_objects 20000 Max. number of objects in the buffer |
rsdb/obj/large_object_size 8192 Bytes Estimation for the size of the largest object |
rsdb/obj/mutex_n 0 Number of mutexes in Export/Import buffer |
rsdb/otr/buffersize_kb 4096 kB Size of OTR buffer |
rsdb/otr/max_objects 2000 Max. number of objects in the buffer |
rsdb/otr/mutex_n 0 Number of mutexes in OTR buffer |
rsdb/esm/buffersize_kb 4096 kB Size of exp/imp SHM buffer |
rsdb/esm/max_objects 2000 Max. number of objects in the buffer |
rsdb/esm/large_object_size 8192 Bytes Estimation for the size of the largest object |
rsdb/esm/mutex_n 0 Number of mutexes in Exp/Imp SHM buffer |
rsdb/ntab/entrycount 29989 Max. number of table definitions buffered |
rsdb/ntab/ftabsize 60004 kB Size of field description buffer |
rsdb/ntab/entrycount 29989 Max. number / 2 of table descriptions buffered |
rsdb/ntab/irbdsize 8000 kB Size of initial record buffer |
rsdb/ntab/entrycount 29989 Max. number / 2 of initial records buffered |
rsdb/ntab/sntabsize 3000 kB Size of short nametab |
rsdb/ntab/entrycount 29989 Max. number / 2 of entries buffered |
zcsa/calendar_area 500000 Byte Size of calendar buffer |
zcsa/calendar_ids 200 Max. number of directory entries |
ztta/roll_area 3000320 Byte Roll area per workprocess (total) |
ztta/roll_first 1024 Byte First amount of roll area used in a dialog WP |
ztta/short_area 4000000 Byte Short area per workprocess |
rdisp/ROLL_SHM 32768 8 kB Part of roll file in shared memory |
rdisp/PG_SHM 16384 8 kB Part of paging file in shared memory |
rdisp/PG_LOCAL 150 8 kB Paging buffer per workprocess |
em/initial_size_MB 5600 MB Initial size of extended memory |
em/blocksize_KB 4096 kB Size of one extended memory block |
em/address_space_MB 4096 MB Address space reserved for ext. mem. (NT only) |
ztta/roll_extension 2000000000 Byte Max. extended mem. per session (external mode) |
abap/heap_area_dia 2000000000 Byte Max. heap memory for dialog workprocesses |
abap/heap_area_nondia 0 Byte Max. heap memory for non-dialog workprocesses |
abap/heap_area_total 5872025600 Byte Max. usable heap memory |
abap/heaplimit 40000000 Byte Workprocess restart limit of heap memory |
abap/use_paging 1 Paging for flat tables used (1) or not (0) |
Please give the recommendation as our core operations are affecting due to the slow response.
Regards
Hi Anwer,
Filip is correct, you should never have to restart production servers, especially for performance reasons.
You say "The 25 users are professional users which are working on the VA01 / VF01 / VL01n / VL02n. These transaction codes are utilizing the lot of resources of CPU, Memory and DB, and they are complaining the slow response of the server while others users are execute any processing for the reports or query."
Are all of the above users on the same application server? If so, this can explain some of the performance issues due to the constant buffer swapping from different user types. One suggestion is to use logon groups for each user type to separate the users to different application servers. This way <i>both</i> user groups will experience increased performance.
Best Regards,
Matt
User | Count |
---|---|
81 | |
10 | |
10 | |
9 | |
7 | |
6 | |
6 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.