Re: pg_restore crashes passing NULL to strcmp (9.4 rc1) - Mailing list pgsql-bugs

From Heikki Linnakangas
Subject Re: pg_restore crashes passing NULL to strcmp (9.4 rc1)
Date
Msg-id 54856EAB.9010407@vmware.com
Whole thread Raw
In response to pg_restore crashes passing NULL to strcmp (9.4 rc1)  (Justin Pryzby <pryzby@telsasoft.com>)
Responses Re: pg_restore crashes passing NULL to strcmp (9.4 rc1)
List pgsql-bugs
On 12/06/2014 08:11 AM, Justin Pryzby wrote:
> I'm not sure what's wrong with this dump file; I remember copying it with rsync
> some months ago, experiencing the crash, and recopying it, not able to break
> off and go into debugging postgres at that point.
>
> $ dpkg-query -W postgresql-client-9.4
> postgresql-client-9.4   9.4~rc1-1
>
> (gdb) r
> Starting program: /usr/lib/postgresql/9.4/bin/pg_restore /home/pryzbyj/clientdb/ptci.d/ptci.segv
> [Thread debugging using libthread_db enabled]
> Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
>
> Program received signal SIGSEGV, Segmentation fault.
> ReadToc (AH=AH@entry=0x55555577b3b0)
>      at /build/buildd/postgresql-9.4-9.4~rc1/build/../src/bin/pg_dump/pg_backup_archiver.c:2426
> 2426    /build/buildd/postgresql-9.4-9.4~rc1/build/../src/bin/pg_dump/pg_backup_archiver.c: No such file or
directory.
>
> 2422                 te->owner = ReadStr(AH);
> 2423                 if (AH->version >= K_VERS_1_9)
> 2424                 {
> 2425                         if (strcmp(ReadStr(AH), "true") == 0)
> 2426                                 te->withOids = true;
> 2427                         else
> 2428                                 te->withOids = false;
> 2429                 }
> 2430                 else
> 2431                         te->withOids = true;

Hmm. It's a bit hard to believe that the "te->withOids = true"
assignment could've segfaulted. So the line number probably isn't 100%
accurate.

A corrupted dump file can certainly cause a segfault in pg_restore, so
the dump file was probably somehow damaged while copying it with rsync.
Can't say more than that, I'm afraid, without the dump file itself.

> (gdb) p ReadStr(AH)
> $1 = 0x0

That tries reads the *next* string from the dump file, so it's not very
surprising that that fails. The previous ReadStr() call probably failed
already, and it would be more interesting to to see what went wrong there.

- Heikki

pgsql-bugs by date:

Previous
From: Francisco Olarte
Date:
Subject: Re: BUG #11986: psql uses pager inside Emacs shell buffer (not a terminal)
Next
From: Ronan Dunklau
Date:
Subject: Dumping database creation options and ACLs