On Mon, Jul 25, 2011 at 10:56 AM, Yan Chunlu <springrider@gmail.com> wrote:
> I think the problem is still "invalid record length" and "invalid
> magic number", it start showing right after I complete sync data and
> start slave. If I stop slave later and restart, yes it could show
> xlog not found and can not catch master. but why the "invalid" things
> in the first place?
You might have the same problem which was reported before.
http://archives.postgresql.org/pgsql-hackers/2011-06/msg00661.php
That problem was fixed, and the fix will be included in next minor
update (i.e., 9.0.5).
http://archives.postgresql.org/pgsql-committers/2011-06/msg00101.php
Of course, you can avoid the problem by building PostgreSQL with
gcc != 4.6.0, I think.
Regards,
--
Fujii Masao
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center