Re: postgres startup failure - Mailing list pgsql-general

From Matt Bartolome
Subject Re: postgres startup failure
Date
Msg-id AANLkTimoPWT6zznS9SpotstBKqH3CMVdBSXpyEb9pCm5@mail.gmail.com
Whole thread Raw
In response to Re: postgres startup failure  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: postgres startup failure  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-general
Hi Tom,

On Thu, May 20, 2010 at 11:33 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
Matt Bartolome <mattxbart@gmail.com> writes:
> I'm attempting to start postgres on a standby machine from a backup. Both
> the primary and standby are running postgres 8.4, fedora 12.

8.4.what exactly?
 
I'm running 8.4.2 on the primary and 8.4.4 on the standby via yum. I wouldn't think a minor release change would cause what I'm seeing... I've also built 8.4.2 from source on the standby and I get the same debug output. I should also mention these are both VM's (vmware) and are running under a windows host OS.

> DEBUG:  invoking IpcMemoryCreate(size=32595968)
> DEBUG:  max_safe_fds = 984, usable_fds = 1000, already_open = 6
> DEBUG:  logger shutting down
> DEBUG:  shmem_exit(0): 0 callbacks to make
> DEBUG:  proc_exit(0): 0 callbacks to make
> DEBUG:  exit(0)
> DEBUG:  shmem_exit(-1): 0 callbacks to make
> DEBUG:  proc_exit(-1): 0 callbacks to make

Hmm.  I'm thinking that the postmaster must have simply crashed at some
time between starting the logger subprocess and starting the startup
subprocess.  You could try running the postmaster under gdb to see
if you can get a stack trace, along the lines of

       $ gdb /usr/bin/postgres
       gdb> run -d 3 -D /data/postgres/
       ... crash
       gdb> bt
       ... printout
       gdb> quit

 
gdb output...

DEBUG:  -----------------------------------------
DEBUG:  invoking IpcMemoryCreate(size=32595968)
DEBUG:  max_safe_fds = 980, usable_fds = 1000, already_open = 10
Detaching after fork from child process 18310.
Detaching after fork from child process 18311.

Program exited with code 01.
DEBUG:  logger shutting down
DEBUG:  shmem_exit(0): 0 callbacks to make
DEBUG:  proc_exit(0): 0 callbacks to make
DEBUG:  exit(0)
DEBUG:  shmem_exit(-1): 0 callbacks to make
DEBUG:  proc_exit(-1): 0 callbacks to make
(gdb) bt
No stack.


You might need to "continue" if it decides to stop at the point of
forking off the logger.

                       regards, tom lane

Do you have any other suggestions?

Thank you,
Matt

pgsql-general by date:

Previous
From: Scott Marlowe
Date:
Subject: Re: PostgreSQL database and other daemons crashing
Next
From: Alvaro Herrera
Date:
Subject: Re: Postgres stats collector showing high disk I/O