On 07/16/2011 06:33 PM, Jeff Janes wrote:
2.6 ms for an fsync seems awfully quick. I wonder if EBS uses nonvolatile/battery-backed write cache, or if it just lies about fsync actually hitting disk.
They have the right type of cache in there to make fsync quick, when you happen to be the lucky one to find it free of a write backlog. So the best case is much better than a typical spinning drive with no such cache. The worst case is in the 100ms+ range though on EBS.
-- Greg Smith 2ndQuadrant US greg@xxxxxxxxxxxxxxx Baltimore, MD -- Sent via pgsql-performance mailing list (pgsql-performance@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-performance
