Selena Deckelmann
Topic: 
Linux Filesystem Performance for Databases: What Assumptions Make

Selena Deckelmann works for End Point Corporation from her home in Portland, OR, where she is a PostgreSQL and Perl consultant. She is part of a volunteer group based in Portland that is reviving benchmark testing for PostgreSQL. The group is made up of students, professors and professionals in the area, all interested in database technology. She is the User Group liaison for the PostgreSQL Global Development Group. She currently leads PDXPUG, a PostgreSQL Users Group, and has helped start a programming group, Code-n-Splode, whose goal is to get more women involved in open source. She is also program chair of a grass-roots open source conference, Open Source Bridge (www.opensourcebridge.org). In her spare time, she collects eggs from her chickens, gardens and occasionally mixes drinks for her local Perl Mongers group.

Abstract: 

The Linux operating system provides a number of file systems that can be used, as well as volume management and hardware or software RAID. We are running performance benchmarks for database tuning, and are curious if the file systems really behave like we expect them to, especially when used in conjunction with RAID or volume management. Are these file systems being used in manners for which they were designed? There is also more to file systems than how fast we can read to them or how fast we can write to them. How reliable is the file system, and how do be prove it? We have collected data and will have a server available for development during the conference.

The talk will include data from recent tests that are Postgres specific - focusing hopefully on default_statistics_target and checkpoint_timeout. We're using DBT-2, DBT-3 and TPC-H benchmarks, originally developed at the OSDL by Mark Wong, among others.