Hi Ioana,
I am a little bit confused about the time displayed in
statistics. I have a stored procedure that I execute
twice through sequoia and for the second execution of
the procedure the time in statistics and recovery log
it seems to be much smaller than the time displayed in
the postgres log! Can somebody tell me what I am
missing here?
Yes, the execution time you see in Sequoia is the time perceived by the
client but there are cache effects and asynchronous execution that
affects execution time.
Recovery log updates are fully asynchronous meaning that recovery log
updates will never be accounted in query execution time. Also the
metadatache will prevent re-fetching ResultSet metadata (very slow on
Postgres) when the query is executed for the 2nd time.
Most of the query execution time that appears in your Postgres log is
due to the recovery log but this happens asynchronously (in parallel)
while the result is being return to the client.
Hope this clarifies things,
Emmanuel
--
Emmanuel Cecchet
Chief Scientific Officer, Continuent
Blog: http://emanux.blogspot.com/
Open source: http://www.continuent.org
Corporate: http://www.continuent.com
Skype: emmanuel_cecchet
Cell: +33 687 342 685
_______________________________________________
Sequoia mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/sequoia
