Try as I might, I wasn't able to get a JIT debugger give me a memory
dump. It seems like postgres.exe is not really crashing in the
"unhandled exception" sense (see gdb log below)? Am I missing a
configure option?
(As an aside, what's the best way to get a core dump on Windows? Can
gdb read memory dumps produced by NTSD?)
I was able to attach to the running backend using gdb. That worked
(kind of, see log below).
I ended up modifying the elog again with the following results:
PANIC: right sibling is not next child in "Panel_pkey", parent is 271,
target is 635, rightsib is 629, nextoffset is 87
C:\WINDOWS\system32>C:\mingw\bin\gdb C:\pgsql\bin\postgres.exe
... snip ...
(gdb) attach 2084
Attaching to program `C:\pgsql\bin\postgres.exe', process 2084
[Switching to thread 2084.0x930]
(gdb) break nbtpage.c:983
Breakpoint 1 at 0x41f042: file nbtpage.c, line 983.
(gdb) continue
Continuing.
[Switching to thread 2084.0x81c]
Breakpoint 1, _bt_pagedel (rel=0x196bbb0, buf=1089, vacuum_full=0
'\0')
at nbtpage.c:983
983 nbtpage.c: No such file or directory.
in nbtpage.c
(gdb) print parent
$1 = 271
(gdb) print target
$2 = 635
(gdb) print rightsib
No symbol "rightsib" in current context.
(gdb) print nextoffset
$3 = 87
(gdb) print leftsib
$4 = 636
(gdb) print rightsib
No symbol "rightsib" in current context.
(gdb) continue
Continuing.
Program exited with code 03.
(gdb)
Pete
>>> Tom Lane <tgl@sss.pgh.pa.us> 04/11/06 9:19 pm >>>
"Peter Brant" <Peter.Brant@wicourts.gov> writes:
> PANIC: right sibling is not next child in "Panel_pkey", parent is
271
Hmm ... that's not actually enough info to tell us where to look,
is it :-(. Please add the following variables to the elog message, or
gdb for them if you can:
target
rightsib
nextoffset
regards, tom lane
---------------------------(end of
broadcast)---------------------------
TIP 6: explain analyze is your friend