Thread: pgsql: Prevent autovacuum transactions from running in serializable mod
Prevent autovacuum transactions from running in serializable mode. Force the transaction isolation level to READ COMMITTED in autovacuum worker and launcher processes. There is no benefit to using a higher isolation level, and doing so could result in delaying foreground transactions (or maybe even causing unnecessary serialization failures?). Noted by Dan Ports. Also, make sure we disable zero_damaged_pages and statement_timeout in the autovac launcher, not only workers. Now that the launcher can run transactions, these settings could affect its behavior, and it seems like the same arguments apply to the launcher as the workers. Branch ------ master Details ------- http://git.postgresql.org/pg/commitdiff/73d1bfd0b567ef9a4439be1510c12c346a064d9e Modified Files -------------- src/backend/postmaster/autovacuum.c | 32 +++++++++++++++++++++++++++++++- 1 files changed, 31 insertions(+), 1 deletions(-)
Re: pgsql: Prevent autovacuum transactions from running in serializable mod
From
Andrew Dunstan
Date:
On 11/29/2011 10:40 PM, Tom Lane wrote: > Prevent autovacuum transactions from running in serializable mode. > > Force the transaction isolation level to READ COMMITTED in autovacuum > worker and launcher processes. There is no benefit to using a higher > isolation level, and doing so could result in delaying foreground > transactions (or maybe even causing unnecessary serialization failures?). > Noted by Dan Ports. > > Also, make sure we disable zero_damaged_pages and statement_timeout in > the autovac launcher, not only workers. Now that the launcher can run > transactions, these settings could affect its behavior, and it seems > like the same arguments apply to the launcher as the workers. > > Looks like this broke the buildfarm. cheers andrew
Re: pgsql: Prevent autovacuum transactions from running in serializable mod
From
Andrew Dunstan
Date:
On 11/29/2011 11:03 PM, Andrew Dunstan wrote: > > > > Looks like this broke the buildfarm. > > Actually, it was the pg_dump cleanup. I clicked the wrong message. cheers andrew
Andrew Dunstan <andrew.dunstan@pgexperts.com> writes: > On 11/29/2011 11:03 PM, Andrew Dunstan wrote: >> Looks like this broke the buildfarm. > Actually, it was the pg_dump cleanup. Huh, my compiler didn't complain. Will fix, thanks! regards, tom lane