Re: [HACKERS] Allow to specify (auto-)vacuum cost limits relative to the database/cluster size?

2016-02-24 Thread Joe Conway
On 02/24/2016 08:54 AM, Alvaro Herrera wrote: > Joe Conway wrote: > >> In my experience it is almost always best to run autovacuum very often >> and very aggressively. That generally means tuning scaling factor and >> thresholds as well, such that there are never more than say 50-100k dead >>

Re: [HACKERS] Allow to specify (auto-)vacuum cost limits relative to the database/cluster size?

2016-02-24 Thread Alvaro Herrera
Joe Conway wrote: > In my experience it is almost always best to run autovacuum very often > and very aggressively. That generally means tuning scaling factor and > thresholds as well, such that there are never more than say 50-100k dead > rows. Then running vacuum with no delays or limits runs

Re: [HACKERS] Allow to specify (auto-)vacuum cost limits relative to the database/cluster size?

2016-02-24 Thread Joe Conway
On 02/23/2016 10:23 PM, Robert Haas wrote: > On Tue, Jan 12, 2016 at 6:12 PM, Andres Freund wrote: >> right now the defaults for autovacuum cost limiting are so low that they >> regularly cause problems for our users. It's not exactly obvious that >> any installation above a

Re: [HACKERS] Allow to specify (auto-)vacuum cost limits relative to the database/cluster size?

2016-02-23 Thread Robert Haas
On Tue, Jan 12, 2016 at 6:12 PM, Andres Freund wrote: > right now the defaults for autovacuum cost limiting are so low that they > regularly cause problems for our users. It's not exactly obvious that > any installation above a couple gigabytes definitely needs to change >

Re: [HACKERS] Allow to specify (auto-)vacuum cost limits relative to the database/cluster size?

2016-02-21 Thread Jim Nasby
On 1/12/16 6:42 AM, Andres Freund wrote: Somehow computing the speed in relation to the cluster/database size is probably possible, but I wonder how we can do so without constantly re-computing something relatively expensive? ISTM relpages would probably be good enough for this, if we take the

[HACKERS] Allow to specify (auto-)vacuum cost limits relative to the database/cluster size?

2016-01-12 Thread Andres Freund
Hi, right now the defaults for autovacuum cost limiting are so low that they regularly cause problems for our users. It's not exactly obvious that any installation above a couple gigabytes definitely needs to change autovacuum_vacuum_cost_delay & autovacuum_vacuum_cost_limit/vacuum_cost_limit.