Re: [GENERAL] Buffers: shared hit/read to shared_buffers dependence

2015-09-09 Thread Guillaume Lelarge
2015-09-09 17:06 GMT+02:00 Pavel Suderevsky : > Laurenz, Merlin, > > Thanks a lot for your explanations. > > >Even if postgres does not cache the table, the o/s will probably > > still cache it assuming it has the memory to do so. > > Could you please clarify, do I understand right that there are

Re: [GENERAL] Buffers: shared hit/read to shared_buffers dependence

2015-09-09 Thread Pavel Suderevsky
Laurenz, Merlin, Thanks a lot for your explanations. >Even if postgres does not cache the table, the o/s will probably > still cache it assuming it has the memory to do so. Could you please clarify, do I understand right that there are no way to determine with 'explain' whether postgres applies

Re: [GENERAL] Buffers: shared hit/read to shared_buffers dependence

2015-09-08 Thread Merlin Moncure
On Sat, Sep 5, 2015 at 3:28 PM, Albe Laurenz wrote: > Pavel Suderevsky wrote: >> When I have been passing through "Understanding explain" manual >> (http://www.dalibo.org/_media/understanding_explain.pdf) >> I've faced some strange situation when table with size of 65MB completely >> placed in c

Re: [GENERAL] Buffers: shared hit/read to shared_buffers dependence

2015-09-05 Thread Albe Laurenz
Pavel Suderevsky wrote: > When I have been passing through "Understanding explain" manual > (http://www.dalibo.org/_media/understanding_explain.pdf) > I've faced some strange situation when table with size of 65MB completely > placed in cache with shared_buffers=320MB and it doesn't with shared_b

[GENERAL] Buffers: shared hit/read to shared_buffers dependence

2015-09-04 Thread Pavel Suderevsky
Hi, When I have been passing through "Understanding explain" manual ( http://www.dalibo.org/_media/understanding_explain.pdf) I've faced some strange situation when table with size of 65MB completely placed in cache with shared_buffers=320MB and it doesn't with shared_buffers <= 256MB. Actually be