Re: BUG #8673: Could not open file "pg_multixact/members/xxxx" on slave during hot_standby

On 2013-12-09 17:49:34 +0200, Serge Negodyuck wrote:
> Latest checkpoint's NextXID: 0/90546484
> Latest checkpoint's NextOID: 6079185
> Latest checkpoint's NextMultiXactId: 42049949
> Latest checkpoint's NextMultiOffset: 55384024
> Latest checkpoint's oldestXID: 710
> Latest checkpoint's oldestXID's DB: 1
> Latest checkpoint's oldestActiveXID: 90546475
> Latest checkpoint's oldestMultiXid: 1
> Latest checkpoint's oldestMulti's DB: 1

So, the oldest multi is 1, thus the truncation code wouldn't have been
able to remove any. So I think this really is an independent problem
from the truncation patch. But a problem nonetheless.

> > Could you also provide ls -l pg_multixact/ on both primary and standby?
>
> Did you mean pg_multixact/members/ ?

From both members, and offset. I'd be great if you could attach a ls -lR
of pg_multixact/.

> I't not possible on slave right now. Since I had to re-sync these
> files from master. May be that was not a good idea but it helped.
>
> On master there are files from 0000 to 14078
>
> On slave there were absent files from A1xx to FFFF
> They were the oldest ones. (October, November)

Hm. Were files from before A1xx, I am not sure how to parse your answer.

Greetings,

Andres Freund

--
Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services