Re: Re: Proposed Windows-specific change: Enable crash dumps (like core files)

Magnus Hagander <magnus@hagander.net>

From: Magnus Hagander <magnus@hagander.net>
To: Robert Haas <robertmhaas@gmail.com>
Cc: Tom Lane <tgl@sss.pgh.pa.us>, Craig Ringer <craig@postnewspapers.com.au>, Dave Page <dpage@pgadmin.org>, Heikki Linnakangas <heikki.linnakangas@enterprisedb.com>, pgsql-hackers@postgresql.org
Date: 2010-12-19T16:39:32Z
Lists: pgsql-hackers
On Sun, Dec 19, 2010 at 17:28, Robert Haas <robertmhaas@gmail.com> wrote:
> On Sun, Dec 19, 2010 at 11:07 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> I agree that having the crash dump code know anything specific about the
>> contents of shared memory is a nonstarter --- far too fragile.  But
>> perhaps we could have some simple rule based only on what the kernel
>> knows about the shmem block, like "dump shmem if it's no more than 1GB".
>
> Not sure what knobs we have available, but would there be any value in
> trying to dump the FIRST 1GB?

So such knob that I can find. Basically, we pick a combination of the flags at

http://msdn.microsoft.com/en-us/library/ms680519(v=VS.85).aspx

We could send it as a "user stream", i guess, but it won't be
available in the debugger at the same address space then - just as a
blob of data to process manually. I doubt it's worth it in that
case...

-- 
 Magnus Hagander
 Me: http://www.hagander.net/
 Work: http://www.redpill-linpro.com/