crash with assertions and WAL_DEBUG - Mailing list pgsql-hackers

From Alvaro Herrera
Subject crash with assertions and WAL_DEBUG
Date
Msg-id 20140614205106.GV18688@eldon.alvh.no-ip.org
Whole thread Raw
Responses Re: crash with assertions and WAL_DEBUG  (Alvaro Herrera <alvherre@2ndquadrant.com>)
Re: crash with assertions and WAL_DEBUG  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
I noticed that HEAD crashes at startup with assertions disabled and
WAL_DEBUG turned on:

#2  0x00000000007af987 in ExceptionalCondition (
    conditionName=conditionName@entry=0x974448 "!(CritSectionCount == 0 || (CurrentMemoryContext) == ErrorContext ||
(MyAuxProcType== CheckpointerProcess))",  
    errorType=errorType@entry=0x7e9bec "FailedAssertion",
    fileName=fileName@entry=0x974150 "/pgsql/source/minmax/src/backend/utils/mmgr/mcxt.c",
    lineNumber=lineNumber@entry=670) at /pgsql/source/minmax/src/backend/utils/error/assert.c:54
#3  0x00000000007d2080 in palloc (size=1024) at /pgsql/source/minmax/src/backend/utils/mmgr/mcxt.c:670
#4  0x00000000005f3abe in initStringInfo (str=str@entry=0x7fff11f68a80)
    at /pgsql/source/minmax/src/backend/lib/stringinfo.c:50
#5  0x00000000004f42ca in XLogInsert (rmid=48 '0', rmid@entry=0 '\000', info=<optimized out>,
    info@entry=0 '\000', rdata=rdata@entry=0x7fff11f68d30)
    at /pgsql/source/minmax/src/backend/access/transam/xlog.c:1262
#6  0x00000000004f4f17 in CreateCheckPoint (flags=flags@entry=6)
    at /pgsql/source/minmax/src/backend/access/transam/xlog.c:8197
#7  0x00000000004f8079 in StartupXLOG () at /pgsql/source/minmax/src/backend/access/transam/xlog.c:7097


I'm using an interim fix by excepting the startup process (see attached
patch), like the current code does for checkpointer, but I guess a more
robust solution should be sought.

I find it strange that nobody has seen this before.

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

Attachment

pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: UPDATE SET (a,b,c) = (SELECT ...) versus rules
Next
From: Andres Freund
Date:
Subject: Re: UPDATE SET (a,b,c) = (SELECT ...) versus rules