Re: postgres startup failure

[Matt Bartolome <mattxbart@xxxxxxxxx>]

On Thu, May 20, 2010 at 2:44 PM, Tom Lane <tgl@xxxxxxxxxxxxx> wrote:

Matt Bartolome <mattxbart@xxxxxxxxx> writes:

> 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.

Huh.  So it called exit(1) somewhere, without generating any error
message first.  That's pretty unfriendly.  Try setting a breakpoint
at exit(), and if you can get it to stop there, get a stack trace
from that point.

Dunno if you're familiar with gdb, but the quick way to do this is
to say
       b exit
and then the "run" command.  It's possible that the attempt to set
the breakpoint will fail because libc.so isn't loaded yet, in
which case you need two steps:
       b main
       run ...
       when control stops at main:
       b exit
       continue

Setting the breakpoint (b exit) got me a little farther...

DEBUG:  invoking IpcMemoryCreate(size=32595968)
DEBUG:  max_safe_fds = 980, usable_fds = 1000, already_open = 10
Detaching after fork from child process 19066.
Detaching after fork from child process 19067.

Breakpoint 1, 0x0046efb6 in exit () from /lib/libc.so.6
(gdb) bt
#0  0x0046efb6 in exit () from /lib/libc.so.6
#1  0x082286de in proc_exit ()
#2  0x0820bbdf in ?? ()
#3  0x0820eb63 in ?? ()
#4  <signal handler called>
#5  0x003ec424 in __kernel_vsyscall ()
#6  0x0051433d in ___newselect_nocancel () from /lib/libc.so.6
#7  0x0820d489 in ?? ()
#8  0x08210489 in PostmasterMain ()
#9  0x081b92df in main ()
(gdb) quit
 

                       regards, tom lane

Thanks Tom,
Matt