RE: unexpected heavy buffered IO during partition bitmap index rebuild
- From: "Mark W. Farnham" <mwf@xxxxxxxx>
- To: <JTornblad@xxxxxxxxxx>, <oracle-l@xxxxxxxxxxxxx>
- Date: Wed, 26 Sep 2012 18:39:52 -0400
Ah. I parsed the English differently. I see now that the 35% migration was
in reference to the old segment. That makes sense.
-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx]
On Behalf Of Tornblad, John
Sent: Wednesday, September 26, 2012 6:00 PM
To: oracle-l@xxxxxxxxxxxxx
Subject: RE: unexpected heavy buffered IO during partition bitmap index
rebuild
| I'm confused. After a CTAS, you might still might have chaining, but I
don't
| understand why there would be migration.
Whoops... more precisely stated (I hope)...
"In fact, after CTAS our new recreated segment (the target of the CTAS) is
1/2 the size (in blocks) of the old segment which had 35% of its rows that
had been migrated (the source for the CTAS)."
...by which I was trying to compare the sizes of these two segments: the
"before CTAS" segment (when "slow" under bitmap index rebuild and doing
way--WAY!--more than the necessary amount of IO--even considering migrated
rows... 285 times the size of the segment) and the "after CTAS"
segment (when we had used CTAS to create a new replacement segment for the
partition which definitely made all migrated rows disappear and upon which
building a a bitmap index was nice and fast and incurred the "expected"
amount of IO--namely, reading a total amount of blocks which appears to be
the size of the segment).
-john
</pre>This message is confidential, intended only for the named
recipient(s) and may contain information that is privileged or exempt from
disclosure under applicable law. If you are not the intended recipient(s),
you are notified that the dissemination, distribution, or copying of this
message is strictly prohibited. If you receive this message in error or are
not the named recipient(s), please notify the sender by return email and
delete this message. Thank you.
--
//www.freelists.org/webpage/oracle-l
--
//www.freelists.org/webpage/oracle-l
Other related posts:
