For this one precise release and patch information will be at least useful (if
not essential).
There have been several bugs serially killed where the plans generated for
different parallel children didn’t match up. There was a failure mode where it
just kept trying to come up with a matched plan until it timed out (which could
be very long).
Running with parallel forced to a single node may avoid some of those that were
rooted in different node/instance specific plan differences.
I thought that the last vampire version of this bug had been staked in the
heart circa 2018, but I could be wrong.
mwf
From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On ;
Behalf Of Laurentiu Oprea
Sent: Friday, March 15, 2024 11:58 AM
To: ORACLE-L (oracle-l@xxxxxxxxxxxxx)
Subject: query hanged in parsing
Dear all,
Maybe you can give me an idea, I have a query running in parallel hanging in
parsing phase for very long time, I did a flamegraph and seems like stack is
looking like this (qesmaUpdatePlanSig is in top of flamegraph)
qkaComputePlanSignature
qkaTraverseQknDescendents
qknApplyCallback
qkaPlanSignatureCB
qesmaUpdatePlanSig
Any ideas what could be the reason for this??
Appreciate your help.