Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Bug: Multi-Threaded Server and 60 second delay on queries

Bug: Multi-Threaded Server and 60 second delay on queries

From: aelera <doug.anderson_at_aelera.com>
Date: 2000/07/20
Message-ID: <040278db.f982354c@usw-ex0108-062.remarq.com>#1/1

After migrating to Oracle 8.1.6 on a faster machine we started experiencing 60-second delays before a query would come back to the client. Eventually we found the bugs (1304130 and 1309641) which state that the dispatchers (in MTS) are sometimes going to sleep for 60 seconds instead of returning the results of a query to the client.

To work around this problem until 8.1.7 is released and stable, we have implemented a background job, which performs a type of "PING" to the port that the sleeping dispatcher is listening on. This causes the dispatcher to wake and process anything in its queue.

My process looks for the situation by querying V$QUEUE, calls a free utility, NetCat (nc on WindowsNT), sleeps for 5 seconds, then starts the whole thing over.

If anyone needs more information regarding this particular bug please send an e-mail message to the address below after removing the NOSPAM part.

Received on Thu Jul 20 2000 - 00:00:00 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US