import ABAP running very slow

[Former Member]

guys,

I am installing IDES EHP4 on a solaris box with oracle 10g. The import is running very slow. in 20 hours, it has only imported 10 out of 105 jobs. Top comand shows i have 26 GB free physical memory (out of 48GB total) and also 123 GB free swap out of 150 GB total. CPU load is also only 3 %. I have already configured the project file acording to sap recommedation. MY oracle init file configuration looks like this. Also, you will see my project and user_attr files below. Please advise what i can change tomake things go faster.

###########################################################

1. (c)Copyright SAP AG, Walldorf #

2. @(#) $Id: //bc/701-1_REL/src/ins/SAPINST/impl/tpls/ora/ind/INITSIDBASE.ORA#1 $

###########################################################

###########################################################

1. PART I, TUNING PARAMETERS #

###########################################################

1. SYSTEM GLOBAL AREA BUFFERS

1. Total System Global Area = sum(bytes) of all caches

2. Variable Size = shared_pool_size (approx.)

3. Database Buffers = db_block_buffers * db_block_size

4. Redo Buffers = log_buffer

#

1. unit of shared_pool_size: bytes

2. unit of db_block_buffers: number of cached blocks

3. unit of log_buffer: bytes

shared_pool_size = 2264924160

1. 10% of shared_pool_size

shared_pool_reserved_size = 226492416

log_buffer = 1048576

1. MANAGING SORTS

1. A small sort requires sort_area_retained_size in

2. memory. Larger sorts are allocating segments in

3. PSAPTEMP. They are using sort_area_size in memory.

sort_area_retained_size = 0

sort_area_size = 2097152

#sort_spacemap_size = 512

1. PGA-PARAMETERS

open_cursors = 800

#cursor_space_for_time = true

1. LOCKS

1. dml_locks = processes * 50

dml_locks = 4000

#enqueue_resources = 8000

1. CPU-PARAMETERS

#spin_count = 200

#_cpu_count = 4

1. TUNING ARCHIVING

#log_archive_buffer_size 127

#log_archive_buffers 4

1. USE VECTOR READ

#use_readv = true

1. OTHER TUNING PARAMETERS

#sequence_cache_entries = 100

#sequence_cache_hash_buckets = 89

###########################################################

1. PART II, CHANGEABLE PARAMETERS #

###########################################################

1. CONTROL-FILES

1. the controlfiles should be mirrored to every new

2. database mount point @SAPDATA_HOME@/sapdata<#>

#control_files = XXXXX

1. AUTOMATICLY STARTED BACKGROUND PROCESSES

2. MANAGING LOG SWITCHES

1. checkpoints occur only when switching logs, if

2. log_checkpoint_interval * size of OS blocks > size of

3. the actual redo log

log_checkpoint_interval = 0

#log_checkpoint_timeout = 0

1. PROCESS-PARAMETERS

1. The number of allocated semaphores is equal to the

2. number of processes

3. sessions = 1.2 * processes

processes = 500

sessions = 600

1. AUDITING AND STATISTICS

1. sql_trace=TRUE

2. audit_trail = true

3. db_block_lru_extended_statistics = 1000

4. db_block_lru_statistics = true

###########################################################

1. PART III, STATIC PARAMETERS #

###########################################################

1. DB-NAME

db_name = XXX

1. DB-BLOCKSIZE

db_block_size = 8192

1. DB-FILES

db_files = 254

1. OPTIMIZER MODE

#optimizer_mode = choose

#optimizer_search_limit = 3

1. PATHS / DESTINATIONS / TRACES

1. /xxx\

: trace files of the background

1. processes

2. xxxx

/saptrace/usertrace: trace files of the user processes

1. log_archive_dest is a destination, not a path.

2. The archivefiles get the name

3. XXXXXXX

background_dump_dest = xxxx

user_dump_dest = xxxx

core_dump_dest = /xxxxx

log_archive_dest = xxxxxx

#log_archive_format = %t_%s

1. OTHERx

1. reduce_alarm only supported on HP

#reduce_alarm = TRUE

1. ORACLE OPS PARAMETER

remote_os_authent = true

1. retention time for RMAN backup information in control file

control_file_record_keep_time = 30

#see SAP note 124361

#db_file_multiblock_read_count = 8

log_checkpoints_to_alert = true

1. OPTIMIZER MODE

#optimizer_index_cost_adj = 10

1. AUDITING AND STATISTICS

#timed_statistics = true

compatible = 10.2.0

parallel_execution_message_size = 16384

query_rewrite_enabled = false

replication_dependency_tracking = false

star_transformation_enabled = true

undo_retention = 43200

undo_tablespace = PSAPUNDO

undo_management = AUTO

pga_aggregate_target = 3019898880

sga_max_size = 84529848320

db_cache_size = 4264924160

job_queue_processes = 1

remote_login_passwordfile = exclusive

recyclebin = off

event = "10191 trace name context forever, level 1"

FILESYSTEMIO_OPTIONS = setall

OPTIMPEEK_USER_BINDS = FALSE

BTREE_BITMAP_PLANS = FALSE

INDEXJOIN_ENABLED = FALSE

INMEMORY_UNDO = FALSE

OPTIMIZERMJC_ENABLED = FALSE

SORTELIMINATION_COST_RATIO = 10

TABLELOOKUP_PREFETCH_SIZE = 0

• USER_ATTR FILE --- I have replaced my sid with SID for privacy reasons ************************************

adm::::profiles=Log Management

lp::::profiles=Printer Management

postgres::::type=role;profiles=Postgres Administration,All

root::::project=MYSID;auths=solaris.*,solaris.grant;profiles=Web Console Management,All;lock_after_retries=no;min_label=admin_low;clearance=admin_high

SIDADM::::project=SID

oraSID::::project=SID

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

• PROJECT FILE --- my sid replaced with 'SID'*********************************

system:0::::

user.root:1::::process.max-file-descriptor=(basic,65536,deny),(priv,65536,deny);process.max-sem-nsems=(priv,512,deny);project.max-sem-ids=(priv,512,deny);project.max-shm-ids=(priv,256,deny);project.max-shm-memory=(priv,18446744073709551615,deny)

noproject:2::::

default:3::::process.max-file-descriptor=(basic,65536,deny),(priv,65536,deny);process.max-sem-nsems=(priv,512,deny);project.max-sem-ids=(priv,512,deny);project.max-shm-ids=(priv,256,deny);project.max-shm-memory=(priv,18446744073709551615,deny)

group.staff:10::::

SID:400:SAP System SID:SIDadm,root,oraDBSID::process.max-file-descriptor=(basic,65536,deny),(priv,65536,deny);process.max-sem-nsems=(priv,512,deny);project.max-sem-ids=(priv,512,deny);project.max-shm-ids=(priv,256,deny);project.max-shm-memory=(priv,18446744073709551615,deny)

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

what can i do to make the import faster ? my system definitely has enough resources but i am not able to utilizethem all. Thanks in advance

*FYI - I chose 4 as the parallel processes. I have 12 physical CPUs on my system and each physical CPU has 2 virtual CPUs. Should i restart the installation with more parallel processes. I have read that they can be set to 3-4 * no. of CPUs. should i set itto 48 or more ?*

Edited by: rajivshekhar on May 10, 2010 8:35 PM

Edited by: rajivshekhar on May 10, 2010 8:37 PM

[Former Member]

I would definitely increase the number of parallel processes. Normally I'm using one process per CPU.

Do you see errors in the Oracle alert log, which might indicate wrong parameter settings?

Kind regards,

Mark

[JPReyes]

You can certainly increase the number of parallel processes... I usually use 1 parallel process per CPU if you have 12 dual cores you could use up to 24 parallel processes.

Worthy to give it a shot.

Regards

Juan

[Former Member]

For anyone who faces this problem:

I increased the number of parallel jobs to 40 and the installation of this IDES system completed in about 20 hours -- a lot faster. Thanks Juan

RS