Syncing pg_multixact directories
Hello hackers,
I think we should be ensuring that directory entries for newly created
multixact files are durable at checkpoint time. Please see attached.
Attachments:
0001-Fix-missing-fsync-of-multixact-directories.patchtext/x-patch; charset=US-ASCII; name=0001-Fix-missing-fsync-of-multixact-directories.patchDownload+7-14
Hi,
On 2020-09-23 13:45:51 +1200, Thomas Munro wrote:
I think we should be ensuring that directory entries for newly created
multixact files are durable at checkpoint time. Please see attached.
Good catch! Probably that should probably be backpatched...
Greetings,
Andres Freund
On Tue, Sep 22, 2020 at 07:05:36PM -0700, Andres Freund wrote:
On 2020-09-23 13:45:51 +1200, Thomas Munro wrote:
I think we should be ensuring that directory entries for newly created
multixact files are durable at checkpoint time. Please see attached.Good catch! Probably that should probably be backpatched...
+1. Passing that down to the SLRU layer is a nice thing to do. Were
you planning to send a second patch here? The commit log generated
mentions patch 1/2.
--
MichaelOn Wed, Sep 23, 2020 at 4:09 PM Michael Paquier <michael@paquier.xyz> wrote:
On Tue, Sep 22, 2020 at 07:05:36PM -0700, Andres Freund wrote:
On 2020-09-23 13:45:51 +1200, Thomas Munro wrote:
I think we should be ensuring that directory entries for newly created
multixact files are durable at checkpoint time. Please see attached.Good catch! Probably that should probably be backpatched...
+1. Passing that down to the SLRU layer is a nice thing to do. Were
you planning to send a second patch here? The commit log generated
mentions patch 1/2.
Oh, that's just because I also have another patch, for master only, to
go on top, but that's in another thread about SLRU fsync offloading.
Sorry for the confusion.
On Wed, Sep 23, 2020 at 4:11 PM Thomas Munro <thomas.munro@gmail.com> wrote:
On Wed, Sep 23, 2020 at 4:09 PM Michael Paquier <michael@paquier.xyz> wrote:
On Tue, Sep 22, 2020 at 07:05:36PM -0700, Andres Freund wrote:
Good catch! Probably that should probably be backpatched...
+1. Passing that down to the SLRU layer is a nice thing to do. Were
you planning to send a second patch here? The commit log generated
mentions patch 1/2.
While back-patching I realised that 9.5 and 9.6 had the same problem
for other SLRUs, so I updated the commit message accordingly and
pushed. Thanks!