Re: Postgres, fsync, and OSs (specifically linux)
От | Craig Ringer |
---|---|
Тема | Re: Postgres, fsync, and OSs (specifically linux) |
Дата | |
Msg-id | CAMsr+YGhFKq1LcabtY61YQtweF7bvp7krYZkCJhXw_JAoa=dvg@mail.gmail.com обсуждение исходный текст |
Ответ на | Re: Postgres, fsync, and OSs (specifically linux) (Andres Freund <andres@anarazel.de>) |
Ответы |
Re: Postgres, fsync, and OSs (specifically linux)
|
Список | pgsql-hackers |
On 18 May 2018 at 00:44, Andres Freund <andres@anarazel.de> wrote:
Hi,
On 2018-05-10 09:50:03 +0800, Craig Ringer wrote:
> while ((src = (RewriteMappingFile *) hash_seq_search(&seq_status)) != NULL)
> {
> if (FileSync(src->vfd, WAIT_EVENT_LOGICAL_REWRITE_SYNC) != 0)
> - ereport(ERROR,
> + ereport(PANIC,
> (errcode_for_file_access(),
> errmsg("could not fsync file \"%s\": %m", src->path)));
To me this (and the other callers) doesn't quite look right. First, I
think we should probably be a bit more restrictive about when PANIC
out. It seems like we should PANIC on ENOSPC and EIO, but possibly not
others. Secondly, I think we should centralize the error handling. It
seems likely that we'll acrue some platform specific workarounds, and I
don't want to copy that knowledge everywhere.
Also, don't we need the same on close()?
Yes, we do, and that expands the scope a bit.
I agree with Robert that some sort of filter/macro is wise, though naming it clearly will be tricky.
I'll have a look.
В списке pgsql-hackers по дате отправления: