7.3 pg_dump with -Fc option crashes

Started by Laurette Cisnerosabout 23 years ago8 messages
#1Laurette Cisneros
laurette@nextbus.com

This works:
pg_dump -h myhost -p 5432 -f mydb.cpgdmp mydb

This does not:
pg_dump -h myhost -p 5432 -Fc -f mydb.cpgdmp mydb
Segmentation fault (core dumped)

Nor does this:
pg_dump -h myhost -p 5432 -Ft -f mydb.cpgdmp mydb
(but I need the -Fc badly as my dbs backup up to large files)

Here's a stack track (if needed):
(adb) bt
#0 0x080562a6 in WriteStr (AH=0x8074638,
c=0x6c627570 <Address 0x6c627570 out of bounds>) at
pg_backup_archiver.c:1519
#1 0x080569b2 in WriteToc (AH=0x8074638) at pg_backup_archiver.c:1851
#2 0x080594b2 in _CloseArchive (AH=0x8074638) at pg_backup_custom.c:802
#3 0x080545ef in CloseArchive (AHX=0x8074638) at pg_backup_archiver.c:113
#4 0x0804a846 in main (argc=9, argv=0xbffff4cc) at pg_dump.c:645
#5 0x401ef306 in __libc_start_main (main=0x8049db0 <main>, argc=9,
ubp_av=0xbffff4cc, init=0x8049634 <_init>, fini=0x8064720 <_fini>,
rtld_fini=0x4000d2dc <_dl_fini>, stack_end=0xbffff4bc)
at ../sysdeps/generic/libc-start.c:129

Thanks,

--
Laurette Cisneros
The Database Group
(510) 420-3137
NextBus Information Systems, Inc.
wwy.nextbus.com
----------------------------------
Life is an SQL old chum...

#2Laurette Cisneros
laurette@nextbus.com
In reply to: Laurette Cisneros (#1)
Re: 7.3 pg_dump with -Fc option crashes

Oh goodness it's even worse as pg_restore can't read the archive from the
first pg_dump:

pg_dump -h myhost -p 5432 -f mydb.pgdmp mydb
pg_restore -l mydb.pgdmp

pg_restore: [archiver] input file does not appear to be a valid archive

Thanks,

L.
On Fri, 10 Jan 2003, Laurette Cisneros wrote:

This works:
pg_dump -h myhost -p 5432 -f mydb.cpgdmp mydb

This does not:
pg_dump -h myhost -p 5432 -Fc -f mydb.cpgdmp mydb
Segmentation fault (core dumped)

Nor does this:
pg_dump -h myhost -p 5432 -Ft -f mydb.cpgdmp mydb
(but I need the -Fc badly as my dbs backup up to large files)

Here's a stack track (if needed):
(adb) bt
#0 0x080562a6 in WriteStr (AH=0x8074638,
c=0x6c627570 <Address 0x6c627570 out of bounds>) at
pg_backup_archiver.c:1519
#1 0x080569b2 in WriteToc (AH=0x8074638) at pg_backup_archiver.c:1851
#2 0x080594b2 in _CloseArchive (AH=0x8074638) at pg_backup_custom.c:802
#3 0x080545ef in CloseArchive (AHX=0x8074638) at pg_backup_archiver.c:113
#4 0x0804a846 in main (argc=9, argv=0xbffff4cc) at pg_dump.c:645
#5 0x401ef306 in __libc_start_main (main=0x8049db0 <main>, argc=9,
ubp_av=0xbffff4cc, init=0x8049634 <_init>, fini=0x8064720 <_fini>,
rtld_fini=0x4000d2dc <_dl_fini>, stack_end=0xbffff4bc)
at ../sysdeps/generic/libc-start.c:129

Thanks,

--
Laurette Cisneros
The Database Group
(510) 420-3137
NextBus Information Systems, Inc.
www.nextbus.com
----------------------------------
Life is an SQL old chum...

#3Tom Lane
tgl@sss.pgh.pa.us
In reply to: Laurette Cisneros (#1)
Re: 7.3 pg_dump with -Fc option crashes

Laurette Cisneros <laurette@nextbus.com> writes:

This does not:
pg_dump -h myhost -p 5432 -Fc -f mydb.cpgdmp mydb
Segmentation fault (core dumped)

If you have any comments on views, this is probably an instance of a
known bug:

2002-12-27 12:10 tgl

* src/bin/pg_dump/: pg_dump.c (REL7_3_STABLE), pg_dump.c: Remove
overenthusiastic free'ing of comment dependencies; could lead to
core dump in pg_dump when dumping views having comments. See bug
#855.

This fix postdates 7.3.1, but you could get the patch from the CVS
server or the pgsql-bugs archives.

regards, tom lane

#4Laurette Cisneros
laurette@nextbus.com
In reply to: Tom Lane (#3)
Re: 7.3 pg_dump with -Fc option crashes

Thanks.

I've nulled all the "comment on view"s but still get it...

On Fri, 10 Jan 2003, Tom Lane wrote:

Laurette Cisneros <laurette@nextbus.com> writes:

This does not:
pg_dump -h myhost -p 5432 -Fc -f mydb.cpgdmp mydb
Segmentation fault (core dumped)

If you have any comments on views, this is probably an instance of a
known bug:

2002-12-27 12:10 tgl

* src/bin/pg_dump/: pg_dump.c (REL7_3_STABLE), pg_dump.c: Remove
overenthusiastic free'ing of comment dependencies; could lead to
core dump in pg_dump when dumping views having comments. See bug
#855.

This fix postdates 7.3.1, but you could get the patch from the CVS
server or the pgsql-bugs archives.

regards, tom lane

--
Laurette Cisneros
The Database Group
(510) 420-3137
NextBus Information Systems, Inc.
www.nextbus.com
----------------------------------
Life is an SQL old chum...

#5Laurette Cisneros
laurette@nextbus.com
In reply to: Tom Lane (#3)
Re: 7.3 pg_dump with -Fc option crashes

That did the trick...fixed pg_dump!

And, pg_restore works on it too!

Yay, I can go home now.

Thanks very much for your help!

On Fri, 10 Jan 2003, Tom Lane wrote:

Laurette Cisneros <laurette@nextbus.com> writes:

This does not:
pg_dump -h myhost -p 5432 -Fc -f mydb.cpgdmp mydb
Segmentation fault (core dumped)

If you have any comments on views, this is probably an instance of a
known bug:

2002-12-27 12:10 tgl

* src/bin/pg_dump/: pg_dump.c (REL7_3_STABLE), pg_dump.c: Remove
overenthusiastic free'ing of comment dependencies; could lead to
core dump in pg_dump when dumping views having comments. See bug
#855.

This fix postdates 7.3.1, but you could get the patch from the CVS
server or the pgsql-bugs archives.

regards, tom lane

--
Laurette Cisneros
The Database Group
(510) 420-3137
NextBus Information Systems, Inc.
www.nextbus.com
----------------------------------
Life is an SQL old chum...

#6Bruce Momjian
pgman@candle.pha.pa.us
In reply to: Laurette Cisneros (#5)
Re: 7.3 pg_dump with -Fc option crashes

Are we near needing 7.3.2? The contraint dump patch is post-7.3.1 too,
and Tom applied a big fix recently to 7.3.X.

---------------------------------------------------------------------------

Laurette Cisneros wrote:

That did the trick...fixed pg_dump!

And, pg_restore works on it too!

Yay, I can go home now.

Thanks very much for your help!

On Fri, 10 Jan 2003, Tom Lane wrote:

Laurette Cisneros <laurette@nextbus.com> writes:

This does not:
pg_dump -h myhost -p 5432 -Fc -f mydb.cpgdmp mydb
Segmentation fault (core dumped)

If you have any comments on views, this is probably an instance of a
known bug:

2002-12-27 12:10 tgl

* src/bin/pg_dump/: pg_dump.c (REL7_3_STABLE), pg_dump.c: Remove
overenthusiastic free'ing of comment dependencies; could lead to
core dump in pg_dump when dumping views having comments. See bug
#855.

This fix postdates 7.3.1, but you could get the patch from the CVS
server or the pgsql-bugs archives.

regards, tom lane

--
Laurette Cisneros
The Database Group
(510) 420-3137
NextBus Information Systems, Inc.
www.nextbus.com
----------------------------------
Life is an SQL old chum...

---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to majordomo@postgresql.org so that your
message can get through to the mailing list cleanly

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073
#7Tom Lane
tgl@sss.pgh.pa.us
In reply to: Bruce Momjian (#6)
Re: 7.3 pg_dump with -Fc option crashes

Bruce Momjian <pgman@candle.pha.pa.us> writes:

Are we near needing 7.3.2?

We clearly need to have a 7.3.2, but I was thinking late January would
be about the right time frame. Bugs are still trickling in (eg, the
plpgsql one Neil just identified), and so far we've not seen anything
that would make me feel we need an immediate release ...

regards, tom lane

#8Nathan Mueller
nmueller@cs.wisc.edu
In reply to: Tom Lane (#7)
Re: 7.3 pg_dump with -Fc option crashes

We clearly need to have a 7.3.2, but I was thinking late January would
be about the right time frame. Bugs are still trickling in (eg, the
plpgsql one Neil just identified), and so far we've not seen anything
that would make me feel we need an immediate release ...

I'm biased, but I think the SSL problems fixed since 7.3.1 are pretty
big. 7.3.1 is still pretty much unuseable if SSL support is required.

--Nate