Re: pgsql: Recursively fsync() the data directory after a crash.

On Mon, May 18, 2015 at 12:39 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Bruce Momjian <bruce(at)momjian(dot)us> writes:
>> On Mon, May 18, 2015 at 12:15:03PM -0400, Tom Lane wrote:
>>> It's a trivial enough change that I wouldn't see a problem with doing it
>>> now. But if you do, please get it in by say 2PM EDT.
>
>> Uh, doesn't that change the translated strings? Is that a problem?
>
> Better an untranslated message than a translated one that lacks critical
> info. But anyway, there are likely other instances of that same string
> *with* %m ...

Probably not, because of the way the message is worded. But we could do this:

diff --git a/src/backend/storage/file/fd.c b/src/backend/storage/file/fd.c
index 6fa75d1..bed8478 100644
--- a/src/backend/storage/file/fd.c
+++ b/src/backend/storage/file/fd.c
@@ -2461,7 +2461,7 @@ pre_sync_fname(char *fname, bool isdir)

if (fd < 0)
ereport(FATAL,
- (errmsg("could not open file \"%s\"
before fsync",
+ (errmsg("could not open file \"%s\": %m",
fname)));

pg_flush_data(fd, 0, 0);

Does that sound good?

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company