| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Usenet -> c.d.o.server -> Re: Need HELP with Deadlocks
I will try the trace and get back to this posting. I also have been reading up on deadlocks. We have DISCRETE_TRANSACTIONS_ENABLED=TRUE. Could this be part of the problem since the updates to a single datablock could be many in the case I have given?
TIA
Eric Peterson
In article <8ec73f$tk4$1_at_nnrp1.deja.com>,
Mark D Powell <markp7832_at_my-deja.com> wrote:
> In article <8ec1n6$n4h$1_at_nnrp1.deja.com>,
> miaemp_at_my-deja.com wrote:
> > Is there some tool out there that will parse this trace file, or how
or
> > what do I look for in the trace file to determine the deadlock.
Here
> > is part of the trace file.
> > Deadlock graph:
> > ---------Blocker(s)-------- ---------Waiter
(s)--
> > -------
> > Resource Name process session holds waits process session
> > holds waits
> > TX-001e0013-0000063d 73 116 X 33
> > 87 S
> > TX-000b0076-0000b2d2 33 87 X 73
> > 116 S
> > Rows waited on:
> > Session 87: no row
> > Session 116: no row
> >
> > To me, it appears that a row is not being waited on, it might be
some
> > sort of resource. Anyone have any other ideas?
> >
> > TIA
> > Eric Peterson
> > Programmer/Analyst DBA
> > Maurices Inc.
> > eric_NOpeterson_at_mauricesSPAMMERS.inrg.com
> >
> > Note that the opinions I may have expressed here are solely my own
and
> > not that of my employer's. To email me directly, remove the NO and
the
> > SPAMMERS from the address.
> >
> > In article <8ea8th$q84$1_at_nnrp1.deja.com>,
> > Mark D Powell <markp7832_at_my-deja.com> wrote:
> > > In article <8ea3r2$k2q$1_at_nnrp1.deja.com>,
> > > miaemp_at_my-deja.com wrote:
> > > > I am on 7.3.4.0 with AIX 4.2. Currently, we have a third party
batch
> > > > job which is logically threaded by a location and we are running
> > > > several threads at once. We have been running this job for
approx.
18
> > > > months with no deadlocking problems. I have checked all of the
tables
> > > > it is accessing to make sure that the foreign keys it has are
indexed,
> > > > so, that is not the problem. What we are experiencing are
deadlocks,
> > > > we find that the thread run fine for a short time, then, the
deadlocks
> > > > start to occur. I looked at the SQL each thread is running, all
are
> > > > trying to insert into the same table except for one, the one is
trying
> > > > to select from a different table. I have checked DBA_WAITERS
and
have
> > > > found all the threads waiting. I have checked DBA_blockers and
there
> > > > are no rows returned. The one thread that is doing the select
seems
to
> > > > be blocking all the others, and all the others are blocking the
one
> > > > that is doing the select, I have not been able to figure out
why.
The
> > > > tables that are being inserted and selected, has the freelists
set
to
> > > > 10 and the initrans set to 10 as well, so, I don't think it's a
problem
> > > > of trying to insert into the same block, but, maybe it is.
> > > >
> > > > Anyone have any suggestions as to how to solve this problem? I
would
> > > > greatly appreciate the help.
> > > >
> > > > TIA
> > > > Eric Peterson
> > > > Programmer/Analyst DBA
> > > > Maurices Inc.
> > > > eric_NOpeterson_at_mauricesSPAMMERS.inrgALLOWED.com
> > > >
> > > > Note that the views I may have expressed here are solely my own
and
not
> > > > that of my employer's. To email me directly, remove the NO
SPAMMERS
> > > > ALLOWED from the email address.
> > > >
> > > I believe that Oracle deadlocks cause a trace file to the user
dump
> > > destination and this trace file contains the rowid of the row the
> > > deadlock was on. You may to look for to see if such trace files
exist
> > > and if they are of any use.
> > > --
> > > Mark D. Powell -- The only advice that counts is the advice that
> > > you follow so follow your own advice --
> > >
> I can not find a copy of the post where I read how to find the
> information but since the trace file for the deadlock does exist try
> using tkprof on it just like it was a session trace file and see what
> happens. Please post if this works or fails.
>
>
Sent via Deja.com http://www.deja.com/
Before you buy.
Received on Fri Apr 28 2000 - 00:00:00 CDT
 |
 |