Re: [PERFORM] index fragmentation on insert-only table with non-unique column

On Fri, Jun 3, 2016 at 8:54 PM, Justin Pryzby wrote: > As a test, I did SET effective_cache_size='1MB', before running explain, and > still does: > > |-> Index Scan using > cdrs_huawei_pgwrecord_2016_05_29_recordopeningtime_idx on > cdrs_huawei_pgwrecord_2016_05_29 (cost=0.44..1526689.

Re: [PERFORM] index fragmentation on insert-only table with non-unique column

On Fri, Jun 03, 2016 at 06:26:33PM -0300, Claudio Freire wrote: > On Wed, May 25, 2016 at 11:00 AM, Justin Pryzby wrote: > >> > First, I found I was able to get 30-50min query results on full week's > >> > table by > >> > prefering a seq scan to an index scan. The row estimates seemed fine, > >

[PERFORM] slony rpm help slony1-95-2.2.2-1.rhel6.x86_64

Hi All Can anyone please point me to location from where i can get slony slony1-95-2.2.2-1.rhel5.x86_64 rpm. I'm upgrading database from version 9.3 to 9.5. Current version of rpm we are using is slo

Re: [PERFORM] index fragmentation on insert-only table with non-unique column

On Wed, May 25, 2016 at 11:00 AM, Justin Pryzby wrote: >> > First, I found I was able to get 30-50min query results on full week's >> > table by >> > prefering a seq scan to an index scan. The row estimates seemed fine, and >> > the >> > only condition is the timestamp, so the planner's use of