Date: Tue, 11 Nov 2008 22:09:38 +0800 From: "Ken Chen" <ken73.chen@gmail.com> To: stable@freebsd.org Subject: Re: php-cgi frozen with sbwait when SMP enable Message-ID: <d018a9bd0811110609oa905044o910cc4f3bab85b67@mail.gmail.com> In-Reply-To: <4917285C.2030702@lozenetz.org> References: <d018a9bd0811070329s30b463b0m6885e1f568217b89@mail.gmail.com> <20081107142137.GA7051@icarus.home.lan> <d018a9bd0811070645g796ff0a3l7a46ffba8ecdd4e2@mail.gmail.com> <4917285C.2030702@lozenetz.org>
next in thread | previous in thread | raw e-mail | index | archive | help
I think the parent php-cgi are very health. I have tried:
There are total 49 php-cgi processes are running or frozen, the '1 wait' is
parent .
web4# ps alx | grep php-cgi | grep -v grep | awk '{print $9}' | sort | uniq
-c | sort -n
1 biowr
1 wait
15 sbwait
32 accept
Kill one of frozen php-cgi processes.
web4# kill -9 61392
Check again the amount of php-cgi processes, there are still 49 php-cgi
procerss.
web4# ps alx | grep php-cgi | grep -v grep | awk '{print $9}' | sort | uniq
-c | sort -n
1 biord
1 bo_wwa
1 wait
4 -
17 sbwait
25 accept
2008/11/10 Anton - Valqk <lists@lozenetz.org>
> Oh, just saw that, this could be caused by dead parent php-cgi processes
> (just a guess).
> I used to run lighttpd with span-fcgi executable and it happens very
> often to have dead parents (of php-cgi childs) that must be killed by
> killall php-cgi (eg. restart _ALL_ php-cgi processes, pretty stupid!!!
> but if you have dead parent you can't know which childs to kill)...
> If you run your php-cgi processes just from the lighttpd(and lighttpd
> manages php-cgi processes) try running it with fcgi-spawn and write a
> script to check parents of the php-cgi backends and you'll see if that's
> the cause of having 'hang' phps :(
>
> pls tell me what is it. I'm interested!
>
> cheers,
> valqk.
> Ken Chen wrote:
> > Hi Jeremy,
> >
> > A health FastCGI process have a lifetime, so the PIDs of all php-cgi
> > processes should in a short range.
> >
> > There are some 'php-cgi' fall in 'sbwait' state, and stay there forever.
> The
> > frozen 'php-cgi' can't accept new request, so never retire.
> >
> > Please forgive my poor English.
> >
> > 2008/11/7 Jeremy Chadwick <koitsu@freebsd.org>
> >
> >
> >> On Fri, Nov 07, 2008 at 07:29:37PM +0800, Ken Chen wrote:
> >>
> >>> Hello,
> >>>
> >>> I have 4 web servers with lighttpd to serve one web site with DNS load
> >>> sharing. On the 2 SMP-enable web servers, there will be many php-cgi
> >>>
> >> frozen
> >>
> >>> in 'sbwait' state every day. It means the php-cgi stay in 'sbwait'
> state,
> >>> and never be back to 'accept' or other state. If I restart them, there
> >>>
> >> will
> >>
> >>> be frozen php-cgi appear some hours later.
> >>>
> >>> There is no problem on the other single CPU web servers which running
> >>>
> >> same
> >>
> >>> php scripts and same configuration and version of PHP.
> >>>
> >>> Why and any solution?
> >>>
> >> I'm not understanding what the problem is (and I've seen the output you
> >> provided later in the thread). Are you stating the problem is that you
> >> see many php-cgi processes? Or are you worried they're not doing
> >> anything? Does the website function, lock up, or anything like that?
> >> If not, what's the issue? :-)
> >>
> >> --
> >> | Jeremy Chadwick jdc at parodius.com |
> >> | Parodius Networking http://www.parodius.com/ |
> >> | UNIX Systems Administrator Mountain View, CA, USA |
> >> | Making life hard for others since 1977. PGP: 4BD6C0CB |
> >>
> >>
> >>
> > _______________________________________________
> > freebsd-stable@freebsd.org mailing list
> > http://lists.freebsd.org/mailman/listinfo/freebsd-stable
> > To unsubscribe, send any mail to "freebsd-stable-unsubscribe@freebsd.org
> "
> >
> >
>
>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?d018a9bd0811110609oa905044o910cc4f3bab85b67>