PDQ and 940
Paul Watson reported that on Solaris 2.8 with IDS 9.40.FC1/2 he could not turn off PDQ except in the onconfig, and queries spawned 20+ threads. Art Kagel said threads grabbed by a PDQ'd procedure stay with that session, so he advised retesting the raw SQL in a clean session. Watson was retesting; no outcome given.
Auto-generated by Claude from the posts below — may be imperfect; read the full thread.
Sun Solaris 2.8 64bit 9.40.FC1/2 I can't turn off PDQ unless it's done in the onconfig. Queries that were running under 9.30 now take forever as long and spawn 20+ threads Anybody else seen this -- Paul Watson # Oninit Ltd # Growing old is mandatory Tel: +44 1436 672201 # Growing up is optional Fax: +44 1436 678693 # Mob: +44 7818 003457 # www.oninit.com #
On Mon, 01 Dec 2003 20:28:26 +0000, Paul Watson <paul@oninit.com> wrote: >Sun Solaris 2.8 64bit >9.40.FC1/2 > >I can't turn off PDQ unless it's done in the onconfig. Queries >that were running under 9.30 now take forever as long and spawn >20+ threads > >Anybody else seen this Nothing having to do with update statistics for procedures with PDQ enabled, is it??
No, I don't think so but I can check. Basically no matter what PDQ is set to in the environment or in the SQL the query is parallel. If I rip the SQL out the SPL it's still PDQ'ing John Carlson wrote: > > On Mon, 01 Dec 2003 20:28:26 +0000, Paul Watson <paul@oninit.com> > wrote: > > >Sun Solaris 2.8 64bit > >9.40.FC1/2 > > > >I can't turn off PDQ unless it's done in the onconfig. Queries > >that were running under 9.30 now take forever as long and spawn > >20+ threads > > > >Anybody else seen this > > Nothing having to do with update statistics for procedures with PDQ > enabled, is it?? -- Paul Watson # Oninit Ltd # Growing old is mandatory Tel: +44 1436 672201 # Growing up is optional Fax: +44 1436 678693 # Mob: +44 7818 003457 # www.oninit.com #
On Tue, 02 Dec 2003 06:09:44 -0500, Paul Watson wrote:
> No, I don't think so but I can check. Basically no matter what PDQ is set to
> in the environment or in the SQL the query is parallel.
>
> If I rip the SQL out the SPL it's still PDQ'ing
Are you certain Paul? If you are relying on checking onstat -g ses and you are
running the raw SQL in the same session that already ran the SPL that may have
been update stats'd or created with a non-zero PDQPRIORITY you should know that
the threads that a PDQ'd procedure grab for a session remain with that session
forever! I reported this 'bug' (in my opinion) four years ago and it's lived
on through more than 5 major releases ever since 7.24 at least (though probably
before)! Try the raw SQL in another session that shows a clean onstat -g ses
before you run it and see what it shows then.
Art S. Kagel
> John Carlson wrote:
>>
>> On Mon, 01 Dec 2003 20:28:26 +0000, Paul Watson <paul@oninit.com> wrote:
>>
>> >Sun Solaris 2.8 64bit
>> >9.40.FC1/2
>> >
>> >I can't turn off PDQ unless it's done in the onconfig. Queries that were
>> >running under 9.30 now take forever as long and spawn 20+ threads
>> >
>> >Anybody else seen this
>>
>> Nothing having to do with update statistics for procedures with PDQ enabled,
>> is it??
>
Art
"Art S. Kagel" wrote:
>
> On Tue, 02 Dec 2003 06:09:44 -0500, Paul Watson wrote:
>
> > No, I don't think so but I can check. Basically no matter what PDQ is set to
> > in the environment or in the SQL the query is parallel.
> >
> > If I rip the SQL out the SPL it's still PDQ'ing
>
> Are you certain Paul?
I was but I'm getting it retested.
> If you are relying on checking onstat -g ses and you are
> running the raw SQL in the same session that already ran the SPL that may have
[cutting]
--
Paul Watson #
Oninit Ltd # Growing old is mandatory
Tel: +44 1436 672201 # Growing up is optional
Fax: +44 1436 678693 #
Mob: +44 7818 003457 #
www.oninit.com #
Related threads
- Posting from the Informix-list
- Migrating from IDS 9.40.UC6 to 11.50.UC3
- Ip for a network session
- questions onstat -g