| From: | Decibel! <decibel(at)decibel(dot)org> |
|---|---|
| To: | Simon Riggs <simon(at)2ndquadrant(dot)com> |
| Cc: | Brad Nicholson <bnichols(at)ca(dot)afilias(dot)info>, pgsql-general(at)postgresql(dot)org |
| Subject: | Re: Cluster and MVCC |
| Date: | 2007-08-15 15:56:15 |
| Message-ID: | 20070815155615.GG54135@nasby.net |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
On Fri, Aug 10, 2007 at 06:34:03PM +0100, Simon Riggs wrote:
> On Fri, 2007-08-10 at 10:02 -0400, Brad Nicholson wrote:
> > I just want to confirm that the cluster/MVCC issues are due to
> > transaction visibility. Assuming that no concurrent access is happening
> > to a given table when the cluster command is issued (when takes it
> > visibility snapshot), it is safe to cluster that table. Correct?
>
> Yes, as long as pre-existing transactions do not then access the
> clustered table. If they do, rows they should have seen will now not be
> visible, yet you won't get an error message to say so.
Don't you also need to be in a serialized transaction?
--
Decibel!, aka Jim Nasby decibel(at)decibel(dot)org
EnterpriseDB http://enterprisedb.com 512.569.9461 (cell)
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Bill Moran | 2007-08-15 15:57:38 | Re: Yet Another COUNT(*)...WHERE...question |
| Previous Message | Vivek Khera | 2007-08-15 15:55:37 | Re: Best practice for: ERROR: invalid byte sequence for encoding "UTF8" |