| From: | Jim Nasby <jim(at)nasby(dot)net> |
|---|---|
| To: | Simon Riggs <simon(at)2ndquadrant(dot)com> |
| Cc: | Daniel Farina <daniel(at)heroku(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Merlin Moncure <mmoncure(at)gmail(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Should I implement DROP INDEX CONCURRENTLY? |
| Date: | 2012-01-03 23:11:37 |
| Message-ID: | FBB4EDF8-2CEB-48B3-8F99-A87352A87847@nasby.net |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Jan 3, 2012, at 12:11 PM, Simon Riggs wrote:
> This could well be related to the fact that DropRelFileNodeBuffers()
> does a scan of shared_buffers, which is an O(N) approach no matter the
> size of the index.
>
> On top of that, taking what Robert Haas mentioned on another thread,
> InvalidateBuffer currently calls StretegyFreeBuffer(), which waits for
> an ExclusiveLock on the BufFreelistLock. On a busy system this will be
> heavily contended, so adding blocks to the freelist only if the lock
> is free seems warranted.
Couldn't we just leave the buffers alone? Once an index is dropped and that's pushed out through the catalog then nothing should be trying to access them and they'll eventually just get aged out.
In fact, IIRC the function that scans for buffers actually checks to see if a rel still exists before it returns the buffer...
--
Jim C. Nasby, Database Architect jim(at)nasby(dot)net
512.569.9461 (cell) http://jim.nasby.net
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Brad Davis | 2012-01-03 23:15:39 | Re: [patch] Improve documentation around FreeBSD Kernel Tuning |
| Previous Message | Jim Nasby | 2012-01-03 23:00:15 | Re: 16-bit page checksums for 9.2 |