Re: pg_dump Crashes and core dumps

Lists: pgsql-bugs
From: Shibashish <shiba(at)it(dot)iitb(dot)ac(dot)in>
To: pgsql-bugs(at)postgresql(dot)org
Cc: pgman(at)candle(dot)pha(dot)pa(dot)us
Subject: pg_dump Crashes and core dumps
Date: 2003-04-23 12:00:19
Message-ID: Pine.LNX.4.44.0304231707280.4598-100000@sarathi.it.iitb.ac.in
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-bugs

============================================================================
POSTGRESQL BUG REPORT TEMPLATE
============================================================================

Your name : Shibashish
Your email address : shib(at)postmark(dot)net

System Configuration
---------------------
Architecture (example: Intel Pentium) : Intel Pentium II

Operating System (example: Linux 2.0.26 ELF) : SCO_SV (SCO Open Server
5.0.4)

PostgreSQL version (example: PostgreSQL-7.3): PostgreSQL-7.3.1

Compiler used (example: gcc 2.95.2) : gcc 2.7.2.1

Please enter a FULL description of your problem:
-----------------------------------------------
I have 2 databases. "test" and "working"
There r 39 tables in "test", very less data in them.
There r 40 tables in "working" and there is some data, not too much. And
there are referential integrety ocnstraints between the data.

When i do a "pg_dump test" it works fine.
But When i do a "pg_dump working" I get "Segmentation fault (core dumped)"

If anybody wants, i can send the core file. Any help will be highly
appreciated.


From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Shibashish <shiba(at)it(dot)iitb(dot)ac(dot)in>
Cc: pgsql-bugs(at)postgresql(dot)org, pgman(at)candle(dot)pha(dot)pa(dot)us
Subject: Re: pg_dump Crashes and core dumps
Date: 2003-04-25 20:26:40
Message-ID: 29730.1051302400@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-bugs

Shibashish <shiba(at)it(dot)iitb(dot)ac(dot)in> writes:
> When i do a "pg_dump test" it works fine.
> But When i do a "pg_dump working" I get "Segmentation fault (core dumped)"
> If anybody wants, i can send the core file. Any help will be highly
> appreciated.

The core file will be useless away from your own system.

Could we see a stack backtrace from it, instead? (Please compile with
debug symbols if you didn't already.) Also, please run pg_dump with
query logging enabled (see log_statement in postgresql.conf) so that you
can show us the last few queries it issues before crashing.

regards, tom lane


From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Shibashish <shiba(at)it(dot)iitb(dot)ac(dot)in>
Cc: pgsql-bugs(at)postgresql(dot)org, pgman(at)candle(dot)pha(dot)pa(dot)us
Subject: Re: pg_dump Crashes and core dumps
Date: 2003-05-03 22:29:30
Message-ID: 16927.1052000970@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-bugs

Shibashish <shiba(at)it(dot)iitb(dot)ac(dot)in> writes:
> When i do a "pg_dump test" it works fine.
> But When i do a "pg_dump working" I get "Segmentation fault (core dumped)"

Olivier Prenant just found a problem that could lead to pg_dump trouble
with any database using PL languages. Try the attached patch and see if
it fixes your problem.

regards, tom lane

*** src/bin/pg_dump/pg_dump.c.orig Thu Apr 24 22:28:22 2003
--- src/bin/pg_dump/pg_dump.c Sat May 3 18:09:51 2003
***************
*** 3591,3597 ****
resetPQExpBuffer(delqry);

/* Make a dependency to ensure function is dumped first */
! deps = malloc(sizeof(char *) * (2 + (strcmp(lanvalidator, "0") != 0) ? 1 : 0));
depIdx = 0;

(*deps)[depIdx++] = strdup(lanplcallfoid);
--- 3591,3597 ----
resetPQExpBuffer(delqry);

/* Make a dependency to ensure function is dumped first */
! deps = malloc(sizeof(char *) * 10);
depIdx = 0;

(*deps)[depIdx++] = strdup(lanplcallfoid);