| Author |
Message
|
| ucbus1 |
 Posted: Thu Aug 01, 2002 8:28 am Post subject: workflow system has become very slow. |
 |
|
Knight
Joined: 30 Jan 2002
Posts: 560
|
We are having workflow system with WINDOWS 2000-AIX setup.The response of the sytem is very... very... slow.
However it is working the way it is supposed to from the application point of view. But the response is very slow.
It has been running quite well for last few months and I am pretty sure that there were no code changes
to the application. We are obserivng this from today morning and any suggestions in this regard are quite useful
Thanks |
|
| Back to top |
|
 |
| jmac |
| Posted: Thu Aug 01, 2002 10:05 am Post subject: |
|
|
Jedi Knight
Joined: 27 Jun 2001
Posts: 3081
Location: EmeriCon, LLC
|
How many instances are "in flight" (i.e running).
How many workItems are there?
Explain a little bit more why it's slow. Is it queries that are slow?, starting activities? etc.
_________________
John McDonald
RETIRED |
|
| Back to top |
|
|
| Ratan |
| Posted: Thu Aug 01, 2002 11:01 am Post subject: |
|
|
Grand Master
Joined: 18 Jul 2002
Posts: 1245
|
How many instances of Execution Server are there. My understanding is you can have 2 instances for each processor on your system to increase performance. Is that true?
-Laze. |
|
| Back to top |
|
|
| ucbus1 |
| Posted: Thu Aug 01, 2002 12:42 pm Post subject: |
|
|
Knight
Joined: 30 Jan 2002
Posts: 560
|
Thanks for the response. Here are more details.
It is slow since the logon itself. It takes few minutes to load the first screen then after 50-60 seconds to load the next screen.
Work Items of worklist (1722)
Process Instances of Process Instance List (842)
We are using windows 2000, webclient 3.3.03, websphere 3.5, IIS 4.0, IE5.5 on client side and on AIX we have workflow 3.3 and DB2 7.1
I am bit puzzled by the sudden slowing down of the process. Checked all the logs for errors.
If I am correct we are having one execution server. Please let me know how I can check the number of servers.
Please let me know if any more information is needed
Thanks |
|
| Back to top |
|
|
| jmac |
| Posted: Thu Aug 01, 2002 1:00 pm Post subject: |
|
|
Jedi Knight
Joined: 27 Jun 2001
Posts: 3081
Location: EmeriCon, LLC
|
You can increase the number of execution servers using the admin utility fmcautil.
See the Admin Guide for details
You really dont have a very heavy load... Are you sure this is not a network problem?
_________________
John McDonald
RETIRED |
|
| Back to top |
|
|
| ucbus1 |
| Posted: Fri Aug 02, 2002 3:31 am Post subject: |
|
|
Knight
Joined: 30 Jan 2002
Posts: 560
|
Is there any criteria when to use multiple execution servers/ restrictions on how many one can use?
Thanks |
|
| Back to top |
|
|
| jmac |
| Posted: Fri Aug 02, 2002 5:11 am Post subject: |
|
|
Jedi Knight
Joined: 27 Jun 2001
Posts: 3081
Location: EmeriCon, LLC
|
I've not seen any doc from IBM here. But realize that adding another execution server on a box that's already running near capacity is not going to help, it will probably hurt.
_________________
John McDonald
RETIRED |
|
| Back to top |
|
|
| manoj |
| Posted: Fri Aug 02, 2002 5:21 am Post subject: |
|
|
Master
Joined: 30 Jan 2002
Posts: 237
Location: Virgina
|
If u r m/c is loaded with other products then it's a good idea not to follow " for every processor u can 've two execution servers". e.g if i 've a 4 processor m/c with following softwares installed MQWorkflow, MQSeries and DB2 then i would keep 4 to 6 execution servers. The point is to keep the number of execution servers less . Again all this depends on m/c also. hope this is useful
regds
manoj |
|
| Back to top |
|
|
| WFUsers |
| Posted: Fri Aug 02, 2002 12:19 pm Post subject: we are having same problem |
|
|
Newbie
Joined: 01 Aug 2002
Posts: 3
|
Hi,
We are having similar problem which is explained by ucbus. we used to have around 30 instances and 60 work items. We increased our application load to around 150 instances and 300 work items. with this load, system response is very slow ( earlier we used to get user worklists in 20 seconds, after load it is taking any where between 50 to 60 seconds ).
Currently we are running 2 execution servers. We checked the m/c cpu utilization, it is around 12%
any ideas ?
Thanks
Surya |
|
| Back to top |
|
|
| kriersd |
| Posted: Sun Aug 04, 2002 4:24 pm Post subject: |
|
|
Master
Joined: 22 Jul 2002
Posts: 209
Location: IA, USA
|
How many workitems are in the worklist your attempting to pull up..
Maybe your filter is not configured to filter out enough workitems.
Just a thought.. Hope it helps.
Dave
_________________
Dave Krier
IBM WebSphere MQ Workflow V3.4 Solution Designer |
|
| Back to top |
|
|
| clindsey |
| Posted: Wed Aug 07, 2002 12:27 pm Post subject: workflow system has become very slow |
|
|
Knight
Joined: 12 Jul 2002
Posts: 586
Location: Dallas, Tx
|
| Running db reorg might help if you haven't done so lately. |
|
| Back to top |
|
|
| WFUsers |
| Posted: Fri Aug 09, 2002 9:35 am Post subject: |
|
|
Newbie
Joined: 01 Aug 2002
Posts: 3
|
I contacted IBm regarding This issue. They suggested some DB2 optmizations. We are in process of implementing these changes. We see little performance improvment.
Thanks
Surya  |
|
| Back to top |
|
|
| kriersd |
| Posted: Tue Oct 01, 2002 9:18 pm Post subject: |
|
|
Master
Joined: 22 Jul 2002
Posts: 209
Location: IA, USA
|
Would you care to share what type of tunning IBM suggested on the DB?
_________________
Dave Krier
IBM WebSphere MQ Workflow V3.4 Solution Designer |
|
| Back to top |
|
|
| kriersd |
| Posted: Sat Oct 05, 2002 8:50 am Post subject: |
|
|
Master
Joined: 22 Jul 2002
Posts: 209
Location: IA, USA
|
Below is information that I posted on another topic. It may help!!!
________
First of all... What platform are you running on? AIX, NT, SUN
Some tips I found.
IN DB2 --> Increase buffer pool from 16 Mb to 128 Mb
IN DB2 --> Decrease audit trail buffer pool from 32 Mb to 8 - 16 Mb
Run runstats, reorgs, and rebinds regularly
Note1: Ensure that you use the fmczutil workflow utility instead of db2
utilities
for the rebind. This will ensure that the correct bind parameters are
used during the bind.
Distribute I/O to several physical disks by spreading out MQ queue, MQ
log, DB log, and DB on separate physical disks
Good Programming practices
Don't query work items or process instances often - obtain a worklist with
all items on it
Don't create work items for high numbers of users - cause a lot of
updates to create and update all the records when a task is worked on or
completed
_________________
Dave Krier
IBM MQSeries
IBM MQSeries Workflow
IBM Websphere
_________________
Dave Krier
IBM WebSphere MQ Workflow V3.4 Solution Designer |
|
| Back to top |
|
|
| amittalekar |
| Posted: Mon Oct 07, 2002 10:23 am Post subject: |
|
|
Disciple
Joined: 03 Apr 2002
Posts: 166
Location: VA, USA
|
Hi,
Some additional DB2 tunning information. Check whether this is helpful ..
1. Monitor Switches
Make sure the monitor switches are turned on. If they aren't,
you won't have access to the performance information you need. To turn the monitor
switches on, issue the command:
db2 "update monitor switches using
lock ON sort ON bufferpool ON uow ON
table ON statement ON"
2. Agents
Make sure there are enough DB2 agents to handle the workload.
To find out, issue the command
db2 "get snapshot for database manager"
and look for the following lines:
High water mark for agents registered = 7
High water mark for agents
waiting for a token = 0
Agents registered= 7
Agents waiting for a token= 0
Idle agents= 5
Agents assigned from pool= 158
Agents created from empty Pool = 7
Agents stolen from another application= 0
High water mark for coordinating agents= 7
Max agents overflow= 0
If you find either Agents waiting for a token or Agents
stolen from another application, increase the number of agents available to the database manager (MAXAGENTS and/or MAX_COORDAGENTS
as appropriate).
3. Maximum Open Files
DB2 has upper limit, on the maximum number of files open at any one time.
The MAXFILOP database configuration parameter stipulates the
maximum number of files that DB2 can have open concurrently. After it reaches that point, DB2 will start closing and opening its
tablespace files (including raw devices). Opening and closing
files slows SQL response times and burns CPU cycles. To find out if DB2 is closing files, issue the command
db2 "get snapshot for database on DBNAME"
and look for the line that reads:
Database files closed = 0
If files are being closed, increase the value of MAXFILOP
until the opening and closing stops. Use the command:
db2 "update db cfg for DBNAME using MAXFILOP N" |
|
| Back to top |
|
|
|
|
