Re: [PERFORM] Yet another abort-early plan disaster on 9.3

2014-12-11 Thread Simon Riggs
On 12 December 2014 at 03:22, Simon Riggs wrote: > It's a simple patch, but it solves the test cases I know about and > does almost nothing to planning time. Test cases attached. The files marked "pettus_*" are written up from Christophe Pettus' blog. The other test case is one of my own devisin

Re: [PERFORM] Yet another abort-early plan disaster on 9.3

2014-12-11 Thread Simon Riggs
On 30 September 2014 at 10:25, Simon Riggs wrote: > On 30 September 2014 00:00, Tom Lane wrote: >> The existing cost estimation >> code effectively assumes that they're perfectly uniformly distributed; >> which is a good average-case assumption but can be horribly wrong in >> the worst case. > >

Re: [PERFORM] Tuning the configuration

2014-12-11 Thread Mark Kirkwood
On 12/12/14 11:36, Eric Pierce wrote: From: pgsql-performance-ow...@postgresql.org on behalf of Evgeniy Shishkin Sent: Thursday, December 11, 2014 7:11 AM To: Andrea Suisani Cc: mfatticci...@mbigroup.it; pgsql-performance@postgresql.org Subject: Re: [P

Re: [PERFORM] Tuning the configuration

2014-12-11 Thread Eric Pierce
From: pgsql-performance-ow...@postgresql.org on behalf of Evgeniy Shishkin Sent: Thursday, December 11, 2014 7:11 AM To: Andrea Suisani Cc: mfatticci...@mbigroup.it; pgsql-performance@postgresql.org Subject: Re: [PERFORM] Tuning the configuration > On

Re: [PERFORM] Tuning the configuration

2014-12-11 Thread Andrea Suisani
On 12/11/2014 01:11 PM, Evgeniy Shishkin wrote: On 11 Dec 2014, at 15:02, Andrea Suisani wrote: On 12/10/2014 11:44 AM, Maila Fatticcioni wrote: 2- I would like to use the two SDD to store the wal file. Do you think it is useful or how should I use them? I definitely would give it a try.

Re: [PERFORM] Tuning the configuration

2014-12-11 Thread Andrea Suisani
Would you mind to explain me better why you do suggest me to use the sas raid for wal please? SSDs are known to shine when they have to deal with random access pattern rather than sequential, on the other hand 10/15K rpm SAS disk is known to be better for sequential io workloads (in general "r

Re: [PERFORM] Tuning the configuration

2014-12-11 Thread Maila Fatticcioni
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 12/11/2014 01:11 PM, Evgeniy Shishkin wrote: > >> On 11 Dec 2014, at 15:02, Andrea Suisani >> wrote: >> >> On 12/10/2014 11:44 AM, Maila Fatticcioni wrote: >>> 2- I would like to use the two SDD to store the wal file. Do >>> you think it is usefu

Re: [PERFORM] Tuning the configuration

2014-12-11 Thread Evgeniy Shishkin
> On 11 Dec 2014, at 15:02, Andrea Suisani wrote: > > On 12/10/2014 11:44 AM, Maila Fatticcioni wrote: >> 2- I would like to use the two SDD to store the wal file. Do you think >> it is useful or how should I use them? > > I definitely would give it a try. > I don't understand the logic behi

Re: [PERFORM] Tuning the configuration

2014-12-11 Thread Andrea Suisani
On 12/10/2014 11:44 AM, Maila Fatticcioni wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hello. I need to tune a postgres installation I've just made to get a better performance. I use two identical servers with a hot replication configuration. The two servers have the following hardware:

Re: [PERFORM] Tuning the configuration

2014-12-11 Thread Maila Fatticcioni
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 12/10/2014 06:47 PM, Patrick Krecker wrote: > On Wed, Dec 10, 2014 at 2:44 AM, Maila Fatticcioni > wrote: Hello. I need to tune a postgres > installation I've just made to get a better performance. I use two > identical servers with a hot replicat