mickyfinn
asked on
DB2 Connect times out after 60 minutes...
I have a query and reporting tool installed on Solaris. We have some very long running queries. Going against Oracle, there seems to be no problem, leading me to believe that the reporting engine (except for potentially the data access libs) is working fine.
When we run any long running query against db2/390, if the join time is over 60 minutes, there is nothing returned ever and the query on the db2 side goes to an RUW waiting state.
On windows, the tool appears to work fine. If the query starts returning data in 50 minutes and this takes longer than 10 minutes, then we're fine. It's only if the client waits for more than 60 minutes for feedback from db2 that we have a problem.
JDBC on the same box apparently has the same problem, (seems to time out, leaves process on host hanging in RUW), but not completely sure.
Does anyone know of a particular idle timeout for the db2 middleware (perhaps particular to DB2/390, not UDB) that can be examined?
Any thing I should look for in db2trc?
Thanks! Micky
When we run any long running query against db2/390, if the join time is over 60 minutes, there is nothing returned ever and the query on the db2 side goes to an RUW waiting state.
On windows, the tool appears to work fine. If the query starts returning data in 50 minutes and this takes longer than 10 minutes, then we're fine. It's only if the client waits for more than 60 minutes for feedback from db2 that we have a problem.
JDBC on the same box apparently has the same problem, (seems to time out, leaves process on host hanging in RUW), but not completely sure.
Does anyone know of a particular idle timeout for the db2 middleware (perhaps particular to DB2/390, not UDB) that can be examined?
Any thing I should look for in db2trc?
Thanks! Micky
ASKER
The problem with it being strictly on the DB2 Server side is that I can run queries against these databases from windows boxes using udbconnect or businessobjects (which uses cli). It appears that this is some sort of setting on the Solaris client side. I had thought it might be a problem with WebIntelligence, but jdbc connections appear to behave similarly. :(
Micky
Micky
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
this maybe what is affecting you...
i believe it just operates on the actual in "SQL" time so if the query only took 59minutes you'd be alright
even if it then took 20minutes to return the results over the network (assuming its set to 60mins)
sorry it been a while since i dealt with this so can't be more specific at present ....
DDS timeout ?
hth