| Author |
Message
|
| PisgahMan |
 Posted: Tue Jul 25, 2006 4:49 am Post subject: SQL Query for clean-up server's work |
 |
|
Voyager
Joined: 27 Jul 2004
Posts: 93
|
We just upgraded to version 3.6 and we are trying to determine exactly how much work clean-up server does and how quickly it does it. Does anyone happen to know the SQL query to determine how many rows there are to be cleaned?
Maybe this? (DB2 database)
select count(*) from fmc.process_inst where state in(??) for read only
We are looking for top level instances, just a rough estimate of how many can be cleaned in what amount of time. |
|
| Back to top |
|
 |
| koko |
| Posted: Tue Jul 25, 2006 6:18 am Post subject: |
|
|
Master
Joined: 26 Sep 2003
Posts: 206
|
| Code: |
| select count(*) from fmc.process_inst where state=128 |
would give you the no.of instances to be physically deleted by the clean up server.
_________________
Thanks
Koko |
|
| Back to top |
|
|
| PisgahMan |
| Posted: Tue Jul 25, 2006 6:42 am Post subject: |
|
|
Voyager
Joined: 27 Jul 2004
Posts: 93
|
Thanks Koko. If in the domain FDL you have
| Code: |
| KEEP_PROCESSES 1 DAYS |
would the query need to include the date as well, or are items in a state of 128 already past that and ready to be deleted? |
|
| Back to top |
|
|
| koko |
| Posted: Tue Jul 25, 2006 7:00 am Post subject: |
|
|
Master
Joined: 26 Sep 2003
Posts: 206
|
The query posted will give you the count of instances which are logically completed in workflow but still present in the runtime database *AT* the time of running the query AFAIK.
_________________
Thanks
Koko |
|
| Back to top |
|
|
| jmac |
| Posted: Tue Jul 25, 2006 3:38 pm Post subject: Re: SQL Query for clean-up server's work |
|
|
Jedi Knight
Joined: 27 Jun 2001
Posts: 3081
Location: EmeriCon, LLC
|
| PisgahMan wrote: |
We just upgraded to version 3.6 and we are trying to determine exactly how much work clean-up server does and how quickly it does it. Does anyone happen to know the SQL query to determine how many rows there are to be cleaned?
Maybe this? (DB2 database)
select count(*) from fmc.process_inst where state in(??) for read only
We are looking for top level instances, just a rough estimate of how many can be cleaned in what amount of time. |
It is my opinion that you need to be very carefull here.... There are a lot more than just process instances to be deleted. The cleanup server will also delete Activity Instances and Block Instances.
_________________
John McDonald
RETIRED |
|
| Back to top |
|
|
| PisgahMan |
| Posted: Wed Jul 26, 2006 11:05 am Post subject: |
|
|
Voyager
Joined: 27 Jul 2004
Posts: 93
|
Thanks guys.
We upgraded the version of Workflow on our production server from 3.4 to 3.6. Usually we run clean up server for 1.5 hours during the night.
After the upgrade the clean-up server filled the RTSINPUTQ to the maxdepth. Exe server could not put anymore messages and locked the system. IBM told us to increase the q maxdepth and let Exe servers catch up, which worked.
After observing the cleanup server run a couple of time, I've noticed it sends a huge amount of messages to the RTSINPUTQ at what seems like the end of its processing. (25,000 on the queue for about 10 minutes) |
|
| Back to top |
|
|
|
|
