Re: 12.3 replicas falling over during WAL redo

Peter Geoghegan wrote on 8/3/20 3:04 PM:

On Mon, Aug 3, 2020 at 2:35 PM Alvaro Herrera <alvherre@2ndquadrant.com> wrote:

You can use pageinspect's page_header() function to obtain the page's
LSN.  You can use dd to obtain the page from the file,

dd if=16605/16613/60529051 bs=8192 count=1 seek=6501 of=/tmp/page.6501

Ben might find this approach to dumping out a single page image
easier, since it doesn't involve relfilenodes or filesystem files:

https://wiki.postgresql.org/wiki/Getting_a_stack_trace_of_a_running_PostgreSQL_backend_on_Linux/BSD#contrib.2Fpageinspect_page_dump

I don't think that works, because it sounds like this requires I connect to the db, but that means I have to start it up. I am (wrongly?) assuming we should get what we can while the db is still in a just-crashed state.

Still, I seem to have mangled Alvaro's suggestion somehow, because this doesn't look right:

# select h.* from public.page , page_header(raw) h;
 lsn | checksum | flags | lower | upper | special | pagesize | version | prune_xid
-----+----------+-------+-------+-------+---------+----------+---------+-----------
 0/0 |        0 |     0 |     0 |     0 |       0 |        0 |       0 |         0
(1 row)

I'll keep trying though.