Hello, I'm very sory to take your time on this mess.
ul 2013 16:06:11 +0530, Amit Kapila amit.kap...@huawei.com wrote in
014201ce7bc6$f71eb950$e55c2bf0$@kap...@huawei.com
I understood your patch's algorithm, but still I have doubt in my mind that
if the next analyze can correct the estimates,
On Wednesday, July 03, 2013 1:21 PM Kyotaro HORIGUCHI wrote:
Hello,
I could see the same output with your latest script, also I could
reproduce
the test if I run the test with individual sql statements.
One of the main point for reproducing individual test was to keep
autovacuum
=
On 07/03/2013 12:51 AM, Kyotaro HORIGUCHI wrote:
It is tough to decide how to modify there. Currently I decided to
preserve vac_estimate_reltuples as possible as it is. For that
objective, I picked up old_rel_tuples as intermediate variable
for the aid to 'deceive' the function. This can be
Hello,
I could see the same output with your latest script, also I could reproduce
the test if I run the test with individual sql statements.
One of the main point for reproducing individual test was to keep autovacuum
= off.
I see. Autovacuum's nap time is 60 sconds for the default
On Thursday, June 27, 2013 4:58 PM Amit Kapila wrote:
On Wednesday, June 26, 2013 7:40 AM Kyotaro HORIGUCHI wrote:
I've recovered from messing up.
snip
Please let me have a bit of time to diagnose this.
I was completely messed up and walking on the wrong way. I looked
into
the
On Wednesday, June 26, 2013 7:40 AM Kyotaro HORIGUCHI wrote:
I've recovered from messing up.
snip
Please let me have a bit of time to diagnose this.
I was completely messed up and walking on the wrong way. I looked into
the vacuum for UPDATEs, not DELETE's so it's quite resonable to have
On Thu, Jun 27, 2013 at 7:27 AM, Amit Kapila amit.kap...@huawei.com wrote:
Now I can look into it further, I have still not gone through in detail
about your new approach to calculate the reltuples, but I am wondering
whether there can be anyway with which estimates can be improved with
Hello,
I have tried to reproduce the problem in different m/c's, but couldn't
reproduce it.
I have ran tests with default configuration.
I think you had reproduced it.
Output on Windows:
---
postgres=# create table t (a int, b int);
(snip)
postgres=# select n_live_tup,
Sorry for lots of mistakes I've made,
I noticed that I was dancing on the 9.3dev at sometime, so I
reran on 9.4devel up to date, having the same result from the
view of inaccuracy of pg_class.reltuples after vacuuming.
Although, what differs from my old testset, vacuum reported that
it sanned
I've recovered from messing up.
snip
Please let me have a bit of time to diagnose this.
I was completely messed up and walking on the wrong way. I looked
into the vacuum for UPDATEs, not DELETE's so it's quite resonable
to have such results.
The renewed test script attached shows the verbose
On Tuesday, June 25, 2013 2:06 PM Kyotaro HORIGUCHI wrote:
Hello,
I have tried to reproduce the problem in different m/c's, but
couldn't
reproduce it.
I have ran tests with default configuration.
I think you had reproduced it.
Output on Windows:
---
postgres=#
On Friday, June 14, 2013 2:05 PM Kyotaro HORIGUCHI wrote:
Hello,
Postgresql estimates the number of live tuples after the vacuum has
left some buffers unscanned. This estimation does well for most cases,
but makes completely different result with a strong imbalance of tuple
density.
For
Sorry, I made an mistake.
Kyotaro HORIGUCHI horiguchi.kyot...@lab.ntt.co.jp:
Overall algorithm could be illistrated as below,
- summing up used bytes, max offnum(PageGetMaxOffsetNumber),
Not max offnum, the number of linp's used after page vacuum.
maximum free bytes for tuple data ,
13 matches
Mail list logo