[BETA7-panic] sodealloc(): so_count 1

[Vlad]

I've installed this patch. Working OK so far, but we don't have that
number of hits coming in lately, so it's hard to tell is it helping or
not. In only crashed for me when it was 60+ requests per second
coming.

I'll keep list posted.

Btw, what's the usual procedure here to commit a certain patch into
-current / -stable tree?  Will it be ever added there?


On Thu, 7 Oct 2004 23:17:38 -0400, Brian Fundakowski Feldman
<green at freebsd.org> wrote:
> On Thu, Oct 07, 2004 at 03:17:57PM -0400, Brian Fundakowski Feldman wrote:
> > On Thu, Oct 07, 2004 at 03:11:48PM -0400, Vlad wrote:
> > > Brian,
> > >
> > > thanks for your work. Question: is this patch is something I can feel
> > > comfortable about trying it on a production server? Did you test it?
> >
> > It is completely untested other than compilation.  I don't have my SMP
> > machine with console in front of me to be able to try to reproduce the
> > problem, but I believe these changes to be relatively safe.
> 
> Okay, now I do; I did a more thorough potential fix -- so if that worked
> for you, this should do the same without a socket memory leak.  The life
> and times of sockets are fraught with peril...
> 
> <URL:http://green.homeunix.org/~green/tcp_accept_race_crash.patch>
> 
> 
> 
> --
> Brian Fundakowski Feldman                           \'[ FreeBSD ]''''''''''\
>   <> green at FreeBSD.org                               \  The Power to Serve! \
>  Opinions expressed are my own.                       \,,,,,,,,,,,,,,,,,,,,,,\
> 


-- 

Vlad