Postgres Pro
Downloads
Home   >   mailing lists

Re: commit dfda6ebaec67 versus wal_keep_segments - Mailing list pgsql-hackers

From Jeff Janes
Subject Re: commit dfda6ebaec67 versus wal_keep_segments
Date
Msg-id CAMkU=1z5doaNOYCdSvH0bDqx29bXrRdKm0SeOw=4iRZMg7GXcA@mail.gmail.com
Whole thread Raw
In response to commit dfda6ebaec67 versus wal_keep_segments  (Jeff Janes <jeff.janes@gmail.com>)
Responses Re: commit dfda6ebaec67 versus wal_keep_segments
List pgsql-hackers
Tree view
On Tue, Apr 2, 2013 at 10:08 PM, Jeff Janes <jeff.janes@gmail.com> wrote:
This commit introduced a problem with wal_keep_segments:

commit dfda6ebaec6763090fb78b458a979b558c50b39b


The problem seems to be that the underflow warned about is happening, because the check to guard it was checking the wrong thing.  However, I don't really understand KeepLogSeg.  It seems like segno, and hence recptr, don't actually serve any purpose.

Cheers,

Jeff

diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 2f9209f..3643be8 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -8264,7 +8264,7 @@ KeepLogSeg(XLogRecPtr recptr, XLogSegNo *logSegNo)
        XLByteToSeg(recptr, segno);

        /* avoid underflow, don't go below 1 */
-       if (segno <= wal_keep_segments)
+       if (*logSegNo <= wal_keep_segments)
                segno = 1;
        else
                segno = *logSegNo - wal_keep_segments;


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Regex with > 32k different chars causes a backend crash
Next
From: Pavel Stehule
Date:
Subject: Re: [PATCH] Exorcise "zero-dimensional" arrays (Was: Re: Should array_length() Return NULL)