Hi,
I suffer from huge increase in number of disk writes after migration
from 8.0.3 to 8.1.2. The average number of sectors written increased 10
times and transfer stays at 20MB of writes per second during work hours
(with 20% cpu time wasted in iowait). The number of queries lasting more
than 300ms increased thrice.
So, any clues what's going on?
The hardware is Dual Opteron 275, 8 GBs RAM, two 73 GB 15K SCSI 320 with
software RAID1. The OS is 64bit Debian Sarge with 8.1.2 build from
sources in exactly the same way as 8.0.3 was built before.
My postgresql.conf:
shared_buffers=30000
max_prepared_transactions=0 #(line added after migration 8.1)
work_mem=256000
maintenance_work_mem=512000
max_fsm_pages=1800000
max_fsm_relations=80000
fsync_on
checkpoints_segments=12
effective_cache_size=800000
log_destination='syslog'
client_min_messages=warning
log_min_messages=notice
log_min_duration_statement=200
stats_start_collector = on
stats_command_string = on
The system is quite busy but the numer of queries hasn't changed after
migration. The cluster contains about 100 databases (some of them with
2000 tables) with total size of 20GBs. During peak hours it hits 160
SELECTS and about 18 modifying queries per second. The DBs are fully
vacuumed every day, and vacuum analyze is performed on critical tables
every 30 minutes.
--
Marcin