await_MC3 ??? How release a thread?

Hi Folks, We migrate a server to 11.50 FC4 on Solaris 9. Some queries which runs on a table stored in 17 dbspaces by round robin with detached indexes are running very slow. If we search only the index it runs well, but if i increase a column who is not in the index, it runs very slow select index where index in (1,2,3..) -> FAST select index, column where index in (1,2,3,...) -> to SLOW. seems like the thread is waiting some resources. looking the output from onstat -g con, this thread (scan_3.1) are waiting for await_MC3 which means "Producer thread(s) are waiting to coordinate their status with their Master Consumer thread" What this really means? how can I manage that? Are there some tasks That i can do to release this thread ? Thanks for you help Miguel Carbone +55 11 96347103 +55 11 78327160 MC IT Solutions Chief Technical Officer miguel@mcsoftware.com.br IIUG - International Informix Users Group Board of Directors miguel@iiug.org --Apple-Mail-416--100589069

The issue is with the introduction of the new non-index column is the select clause is that you move from a key only scan to at best a key-first scan. If the optimizer believes the in clause will scan a large portion of the table then it probably will use a sequential scan. I would make sure to have good distributions on the index column used in the in clause. I would look at the set explain file for both queries. John F. Miller III STSM, Support Architect miller3@us.ibm.com 503-578-5645 IBM Informix Dynamic Server (IDS) ids-bounces@iiug.org wrote on 07/28/2009 06:59:16 AM: > [image removed] > > await_MC3 ??? How release a thread? [16514] > > Miguel Carbone > > to: > > ids > > 07/28/2009 07:00 AM > > Sent by: > > ids-bounces@iiug.org > > Please respond to ids > > Hi Folks, > > We migrate a server to 11.50 FC4 on Solaris 9. > > Some queries which runs on a table stored in 17 dbspaces by round > robin with detached indexes are running very slow. > > If we search only the index it runs well, but if i increase a column > who is not in the index, it runs very slow > > select index where index in (1,2,3..) -> FAST > select index, column where index in (1,2,3,...) -> to SLOW.> > seems like the thread is waiting some resources. > > looking the output from onstat -g con, this thread (scan_3.1) are > waiting for await_MC3 which means "Producer thread(s) are waiting to > coordinate their status with their Master Consumer thread" > > What this really means? how can I manage that? Are there some tasks > That i can do to release this thread ? > > Thanks for you help > > Miguel Carbone > +55 11 96347103 > +55 11 78327160 > > MC IT Solutions > Chief Technical Officer > miguel@mcsoftware.com.br > > IIUG - International Informix Users Group > Board of Directors > miguel@iiug.org > > --Apple-Mail-416--100589069 > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. >