Re: [PERFORM] syslog slowing the database?
> "GS" == Greg Spiegelberg <[EMAIL PROTECTED]> writes: GS> I've been waiting all day for a pg_restore to finish on a test system GS> identically configured as our production in hardware and software GS> with the exception prod is 7.3.5 and test is 7.4.1. GS> The file it's restoring from is about 8GB uncompressed from a GS> "pg_dump -b -F t" and after 2 hours the directory the database is in GS> contains only 1GB. iostat reported ~2000 blocks written every 2 GS> seconds to the DB file system. Have you considered increasing the value of checkpoint_segments to something like 50 or 100 during your restore? It made a *dramatic* improvement on my system when I did the same migration. -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Vivek Khera, Ph.D.Khera Communications, Inc. Internet: [EMAIL PROTECTED] Rockville, MD +1-301-869-4449 x806 AIM: vivekkhera Y!: vivek_khera http://www.khera.org/~vivek/ ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [PERFORM] syslog slowing the database?
On Tuesday 09 March 2004 20:29, Greg Spiegelberg wrote: > iostat reported ~2000 blocks written every 2 > seconds to the DB file system. > > I turned syslog off to see if it was blocking anything and in the > past couple minutes 1GB has been restored and iostat reports ~35,000 > blocks written every 2 seconds to the DB file system. > Can anyone confirm this for me? If syslog is set to sync after every line and you're logging too much then it could slow things down as the disk heads shift back and fore between two areas of disk. How many disks do you have and in what configuration? Also - was PG logging a lot of info, or is some other application the culprit? Tip: put a minus "-" in front of the file-path in your syslog.conf and it won't sync to disk after every entry. -- Richard Huxton Archonet Ltd ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org
[PERFORM] syslog slowing the database?
I've been waiting all day for a pg_restore to finish on a test system identically configured as our production in hardware and software with the exception prod is 7.3.5 and test is 7.4.1. The file it's restoring from is about 8GB uncompressed from a "pg_dump -b -F t" and after 2 hours the directory the database is in contains only 1GB. iostat reported ~2000 blocks written every 2 seconds to the DB file system. I turned syslog off to see if it was blocking anything and in the past couple minutes 1GB has been restored and iostat reports ~35,000 blocks written every 2 seconds to the DB file system. The system is completely idle except for this restore process. Could syslog the culprit? I turned syslog back on and the restore slowed down again. Turned it off and it sped right back up. Can anyone confirm this for me? Greg -- Greg Spiegelberg Sr. Product Development Engineer Cranel, Incorporated. Phone: 614.318.4314 Fax: 614.431.8388 Email: [EMAIL PROTECTED] Cranel. Technology. Integrity. Focus. ---(end of broadcast)--- TIP 8: explain analyze is your friend