> I then did some calculations, based on the sustained write speed of our > drive array (as measured by copying big files to it), and we tried > this: > > bgwriter_lru_percent = 20.0 > bgwriter_lru_maxpages = 200 > bgwriter_all_percent = 10.0 > bgwriter_all_maxpages = 600 > > This almost totally eliminated the clusters of timeouts, and caused the > transaction application rate to increase by a factor of eight over the > already-improved speed. (That is, we were running 30 to 35 times as > many transactions per minute into the database, compared to the default > background writer configuration.) I'm going to let these settings > settle in for a week or two before we try adjusting them further (to see > if we can eliminate those last few timeouts of this type). Can you tell us what type of array you have? Joshua D. Drake > > I guess my point is that people shouldn't be shy about boosting these > numbers by a couple orders of magnitude from the default values. It may > also be worth considering whether the defaults should be something more > aggressive. > > -Kevin > > > ---------------------------(end of broadcast)--------------------------- > TIP 1: if posting/reading through Usenet, please send an appropriate > subscribe-nomail command to majordomo@xxxxxxxxxxxxxx so that your > message can get through to the mailing list cleanly -- The PostgreSQL Company - Command Prompt, Inc. 1.503.667.4564 PostgreSQL Replication, Consulting, Custom Development, 24x7 support Managed Services, Shared and Dedicated Hosting Co-Authors: PLphp, PLperl - http://www.commandprompt.com/
