9.4 checksum error in recovery with btree index - Mailing list pgsql-hackers

From Jeff Janes
Subject 9.4 checksum error in recovery with btree index
Date
Msg-id CAMkU=1yFBE7-anm4saJC4wZ5rkNdhE-3tNfKBHoLA81346zFcg@mail.gmail.com
Whole thread Raw
Responses Re: 9.4 checksum error in recovery with btree index  (Heikki Linnakangas <hlinnakangas@vmware.com>)
List pgsql-hackers

More fun with my torn page injection test program on 9.4.

24171  2014-05-16 14:00:44.934 PDT:WARNING:  01000: page verification failed, calculated checksum 21100 but expected 3356
24171  2014-05-16 14:00:44.934 PDT:CONTEXT:  xlog redo split_l: rel 1663/16384/16405 left 35191, right 35652, next 34666, level 0, firstright 192
24171  2014-05-16 14:00:44.934 PDT:LOCATION:  PageIsVerified, bufpage.c:145
24171  2014-05-16 14:00:44.934 PDT:FATAL:  XX001: invalid page in block 34666 of relation base/16384/16405
24171  2014-05-16 14:00:44.934 PDT:CONTEXT:  xlog redo split_l: rel 1663/16384/16405 left 35191, right 35652, next 34666, level 0, firstright 192
24171  2014-05-16 14:00:44.934 PDT:LOCATION:  ReadBuffer_common, bufmgr.c:483


I've seen this twice now, the checksum failure was both times for the block labelled "next" in the redo record.  Is this another case where the block needs to be reinitialized upon replay?

Cheers,

Jeff

pgsql-hackers by date:

Previous
From: Andres Freund
Date:
Subject: Re: %d in log_line_prefix doesn't work for bg/autovacuum workers
Next
From: Tom Lane
Date:
Subject: Re: %d in log_line_prefix doesn't work for bg/autovacuum workers