| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Kim <kim(at)myemma(dot)com> |
| Cc: | pgsql-performance(at)postgresql(dot)org |
| Subject: | Re: [HACKERS] unusual performance for vac following 8.2 upgrade |
| Date: | 2007-01-11 23:10:20 |
| Message-ID: | 18336.1168557020@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers pgsql-performance |
I wrote:
> What I think we need to do about this is
> (1) fix pgstat_vacuum_tabstats to have non-O(N^2) behavior; I'm thinking
> of using a hash table for the OIDs instead of a linear list. Should be
> a pretty small change; I'll work on it today.
I've applied the attached patch to 8.2 to do the above. Please give it
a try and see how much it helps for you. Some limited testing here
confirms a noticeable improvement in VACUUM startup time at 10000
tables, and of course it should be 100X worse with 100000 tables.
I am still confused why you didn't see the problem in 8.1, though.
This code is just about exactly the same in 8.1. Maybe you changed
your stats collector settings when moving to 8.2?
regards, tom lane
| Attachment | Content-Type | Size |
|---|---|---|
| unknown_filename | text/plain | 5.2 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Simon Riggs | 2007-01-11 23:10:38 | Re: [HACKERS] [PATCHES] wal_checksum = on (default) | off |
| Previous Message | Andrew Dunstan | 2007-01-11 23:04:56 | Re: [HACKERS] Checkpoint request failed on version 8.2.1. |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Kim | 2007-01-11 23:12:46 | Re: unusual performance for vac following 8.2 upgrade |
| Previous Message | Tom Lane | 2007-01-11 22:26:55 | Re: unusual performance for vac following 8.2 upgrade |