Re: [PERFORM] Specifications for a new server

2014-05-06 Thread Dorian Hoxha
And nvram is ram on hardware-raid controller that is not erased on reboot(battery) ? Can this nvram be used also when the configuration is jbod(justabunchofdisks) or some kind of raid(0/1/5/6/10 etc) is required to use the nvram. On Tue, May 6, 2014 at 1:24 PM, Michael Stone wrote: > On Tue, May

Re: [PERFORM] Re: recently and selectively slow, but very simple, update query....

2014-05-06 Thread Stelios Mavromichalis
hello again, some more data on the subject: you are right, delete/re-insert didn't solve the problem (haven't yet tried dump/restore, i might tonight, when low traffic). a short snip from logs were i log all queries that take longer then 1sec: 2014-05-06T15:57:46.303880+03:00 s10 postgres[46220

Re: [PERFORM] Specifications for a new server

2014-05-06 Thread Michael Stone
On Tue, May 06, 2014 at 01:15:10PM +0200, Dorian Hoxha wrote: Since the commitlog/WAL is sequential-write, does it mattert that much to put it in ssd No, assuming a good storage system with nvram write buffer. Mike Stone -- Sent via pgsql-performance mailing list (pgsql-performance@postgr

Re: [PERFORM] Specifications for a new server

2014-05-06 Thread Dorian Hoxha
Since the commitlog/WAL is sequential-write, does it mattert that much to put it in ssd ?(i understand that it matters to put it in separate disk-subsystem so the write/read patterns don't interfere) On Tue, May 6, 2014 at 1:07 PM, Michael Stone wrote: > On Tue, May 06, 2014 at 11:13:42AM +0200,

Re: [PERFORM] Specifications for a new server

2014-05-06 Thread Michael Stone
On Tue, May 06, 2014 at 11:13:42AM +0200, Johann Spies wrote: Analysis or the SAR-logs showed that there were too much iowait in the CPU's on the old system which has a lower spec CPU than the ones considered for the new system. iowait means the cpu is doing nothing but waiting for data from th

Re: [PERFORM] Specifications for a new server

2014-05-06 Thread DFE
I can suggest to have a disks' layout using at least two RAIDs: 1) RAID10 SSD (or 15kRPM HDD) SAS for O.S. and "pg_xlog" folder where PG writes WAL files before checkpoint calls. 2) RAID10 using how many span is possible for the default DB folder. Regards, 2014-05-06 11:13 GMT+02:00 Johann Spi

[PERFORM] Specifications for a new server

2014-05-06 Thread Johann Spies
I am busy reading Gregory Smith' s PostgreSQL 9.0 High Performance and when the book was written he seemed to me a bit sceptical about SSD's. I suspect the reliability of the SSD's has improved significantly since then. Our present server (128Gb RAM and 2.5 Tb disk space and 12 CPU cores - RAID