On Mon, Sep 26, 2016 at 3:44 PM, Tom Lane <[email protected]> wrote:
> Magnus Hagander <[email protected]> writes:
> > Attached patch puts a retry loop around opening the file that retries
> for 5
> > seconds (which is excessive, but should be safe) in case the file is
> > missing (and still fails out for other error messages of course).
>
> > Comments?
>
> The patch assumes that pg_usleep won't change errno, an assumption
> I have little faith in.
>
>
Oh, right, at the very last loop. I've never seen it need more than 1 loop
so I didn't manage to hit that codepath :) But yeah, saving errno and
restoring it on the other side of pg_usleep is probably a good idea.
--
Magnus Hagander
Me: http://www.hagander.net/
Work: http://www.redpill-linpro.com/
diff --git a/src/bin/pg_xlogdump/pg_xlogdump.c b/src/bin/pg_xlogdump/pg_xlogdump.c
index 02575eb..6947c0c 100644
--- a/src/bin/pg_xlogdump/pg_xlogdump.c
+++ b/src/bin/pg_xlogdump/pg_xlogdump.c
@@ -249,6 +249,7 @@ XLogDumpXLogRead(const char *directory, TimeLineID timeline_id,
if (sendFile < 0 || !XLByteInSeg(recptr, sendSegNo))
{
char fname[MAXFNAMELEN];
+ int tries;
/* Switch to another logfile segment */
if (sendFile >= 0)
@@ -258,7 +259,24 @@ XLogDumpXLogRead(const char *directory, TimeLineID timeline_id,
XLogFileName(fname, timeline_id, sendSegNo);
- sendFile = fuzzy_open_file(directory, fname);
+ for (tries = 0; tries < 10; tries++)
+ {
+ sendFile = fuzzy_open_file(directory, fname);
+ if (sendFile >= 0)
+ break;
+ if (errno == ENOENT)
+ {
+ int save_errno = errno;
+
+ /* File not there yet, try again */
+ pg_usleep(500 * 1000);
+
+ errno = save_errno;
+ continue;
+ }
+ /* Any other error, fall through and fail */
+ break;
+ }
if (sendFile < 0)
fatal_error("could not find file \"%s\": %s",
--
Sent via pgsql-hackers mailing list ([email protected])
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
