On Wed, 25 Nov 2009, Grzegorz Jaśkiewicz wrote:
the out of order data layout is primary reason for index bloat. And that
happens , and
gets worse over time once data is more and more distributed. ("random" deletes,
etc).
That's not index bloat. Sure, having the table not in the same order as
Hello,
2009/11/25 Richard Neill :
>It's a simple query, but using a complex view. So I can't really re-order it.
View is inserted directly into your query by PG, and then reordered
according to from_collapse_limit. Probably, problems lies in the view?
How good is it performing? Or from_collapse_l
Hello,
2009/11/25 Richard Neill :
Also, if you find odd statistics of freshly analyzed table - try
increasing statistics target, using
ALTER TABLE .. ALTER COLUMN .. SET STATISTICS ...
If you're using defaults - it's again low for large tables. Start with
200, for example.
Best regards,
Sergey
Sergey Aleynikov wrote:
Hello,
2009/11/25 Richard Neill :
Also, if you find odd statistics of freshly analyzed table - try
increasing statistics target, using
ALTER TABLE .. ALTER COLUMN .. SET STATISTICS ...
If you're using defaults - it's again low for large tables. Start with
200, for exa
Dear All,
I'm wondering whether Vacuum/analyse (notably by the autovaccuum daemon)
is responsible for some deadlocks/dropouts I'm seeing.
One particular table gets hit about 5 times a second (for single row
updates and inserts) + associated index changes. This is a very light
load for the ha
On Thu, Nov 26, 2009 at 4:20 PM, Richard Neill wrote:
> Dear All,
>
> I'm wondering whether Vacuum/analyse (notably by the autovaccuum daemon) is
> responsible for some deadlocks/dropouts I'm seeing.
>
> One particular table gets hit about 5 times a second (for single row
> updates and inserts) +
Richard Neill writes:
> I'm wondering whether Vacuum/analyse (notably by the autovaccuum daemon)
> is responsible for some deadlocks/dropouts I'm seeing.
> One particular table gets hit about 5 times a second (for single row
> updates and inserts) + associated index changes. This is a very ligh
On Thursday 26 November 2009 17:20:35 Richard Neill wrote:
> Dear All,
>
> I'm wondering whether Vacuum/analyse (notably by the autovaccuum daemon)
> is responsible for some deadlocks/dropouts I'm seeing.
>
> One particular table gets hit about 5 times a second (for single row
> updates and inser
Steve Crawford wrote:
> Greg Smith wrote:
>> Jochen Erwied wrote:
>>> - Promise Technology Supertrak ES4650 + additional BBU
>>>
>> I've never seen a Promise controller that had a Linux driver you would
>> want to rely on under any circumstances...
> +1
>
> I haven't tried Promise recently, but
