Re: BUG #7883: "PANIC: WAL contains references to invalid pages" on replica recovery

On Fri, Feb 15, 2013 at 10:57 AM, Maciek Sakrejda <maciek(at)heroku(dot)com> wrote:

> On Fri, Feb 15, 2013 at 7:10 AM, Heikki Linnakangas <
> hlinnakangas(at)vmware(dot)com> wrote:
>
>> Hmm, that sure looks like the same issue Kyotaro HORIGUCHI reported (
>> http://www.postgresql.org/message-id/20121206.130458.170549097.horiguchi.kyotaro@lab.ntt.co.jp),
>> but that was fixed in 9.1.8. Maybe there's some corner case where it's
>> still not working.
>>
>
> We reproduce it on a separate replica with the same base backup. Tried a
> new base backup; that worked.
>
>
>> Did you keep a copy of the WAL involved? Any chance of running xlogdump
>> on it, and posting the results, or just sending over the WAL files so I
>> could take a look?
>>
>
> Yes, we have the WAL. I can't share it as it's customer data. I tried
> setting up xlogdump (this, yes? https://github.com/snaga/xlogdump ), but
> ran into problems building it (tag REL0_5_2):
>
> root(at)host:/home/heroku/xlogdump# PATH=/usr/lib/postgresql/9.1/bin:$PATH
> USE_PGXS=1 make
> gcc -g -O2 -fstack-protector --param=ssp-buffer-size=4 -Wformat
> -Wformat-security -Werror=format-security -fPIC -Wall -Wmissing-prototypes
> -Wpointer-arith -Wdeclaration-after-statement -Wendif-labels
> -Wformat-security -fno-strict-aliasing -fwrapv -fexcess-precision=standard
> -g -I/usr/include/postgresql -DDATADIR=\"/usr/share/postgresql/9.1\" -I.
> -I. -I/usr/include/postgresql/9.1/server -I/usr/include/postgresql/internal
> -D_FORTIFY_SOURCE=2 -DLINUX_OOM_ADJ=0 -D_GNU_SOURCE -I/usr/include/libxml2
> -I/usr/include/tcl8.5 -c -o xlogdump.o xlogdump.c
> xlogdump.c: In function ‘ReadRecord’:
> xlogdump.c:321:7: warning: format ‘%u’ expects argument of type ‘unsigned
> int’, but argument 3 has type ‘long unsigned int’ [-Wformat]
> xlogdump.c: At top level:
> xlogdump.c:865:1: error: conflicting types for ‘ExceptionalCondition’
> /usr/include/postgresql/postgres.h:686:13: note: previous declaration of
> ‘ExceptionalCondition’ was here
> make: *** [xlogdump.o] Error 1
>
>
Can you try master on xlogdump instead of the release? I have fixed a few
bugs there a year ago that don't seem to be in the release, and will fix
more if necessary.

--
fdr