Re: [Ganglia-developers] mod_python

[Brad N. <BNI...@no...>]

  I can check it in.

Brad


>>> On 1/18/2008 at 12:42 PM, in message
<Pin...@pr...>, Matthias
Blankenhaus <mat...@sg...> wrote:

>=20
> On Fri, 18 Jan 2008, Brad Nicholes wrote:
>=20
>>   From what I have read about the popen2 interface, I think you are =
right. =20
> It was either that or having to manually close the stdin.
>>
>=20
> I don't think I haved write access to the tree.  Could you pl. check it =
in=20
> or does Bernard takes care of this ?
>=20
> Matthias
> =20
>> Brad
>>=20
>> >>> On 1/18/2008 at 11:34 AM, in message
>> <Pin...@pr...>, Matthias
>> Blankenhaus <mat...@sg...> wrote:
>>=20
>> >=20
>> > On Fri, 18 Jan 2008, Brad Nicholes wrote:
>> >=20
>> >>   I am seeing the same thing.  It goes away if I use the subprocess =
module.=20
> =20
>> > There must be some problem in the way that popen2 is invoked that =
is=20
> leaving=20
>> > a defunct process.
>> >=20
>> > This fixes the problem:
>> >=20
>> >  diff -u tcpconn.py tcpconn.py.new
>> > --- tcpconn.py  2008-01-18 10:34:05.000000000 -0800
>> > +++ tcpconn.py.new      2008-01-18 10:33:58.000000000 -0800
>> > @@ -212,6 +212,7 @@
>> >              #Call the netstat utility and split the output into =
separate=20
>> > lines
>> >              netstat_output=3Dpopen2.popen2(["netstat", '-t', =
'-a'],=20
>> > mode=3D'r')[0].read()
>> >              lines =3D netstat_output.splitlines()
>> > +            os.wait()
>> >=20
>> >              #Iterate through the netstat output looking for the =
'tcp'=20
>> > keyword in the tcp_at
>> >              # position and the state information in the tcp_state_at=
=20
>> > position. Count each
>> >=20
>> >=20
>> >=20
>> > Matthias
>> >=20
>> >=20
>> >=20
>> >>=20
>> >> Brad
>> >>=20
>> >> >>> On 1/17/2008 at 7:37 PM, in message
>> >> <d4c...@ma...>, =
"Bernard Li"
>> >> <be...@va...> wrote:
>> >> > Hi Brad:
>> >> >=20
>> >> > I'm not sure if my patch caused this, or this is a standard =
behaviour
>> >> > of tcpconn:
>> >> >=20
>> >> > 17825 ?        Ssl    0:00 /usr/sbin/gmond
>> >> > 17864 ?        Z      0:00  \_ [netstat] <defunct>
>> >> >=20
>> >> > Can you check whether you have the same thing when you startup =
your gmond?
>> >> >=20
>> >> > Thanks,
>> >> >=20
>> >> > Bernard
>> >> >=20
>> >> > On 1/16/08, Brad Nicholes <BNI...@no...> wrote:
>> >> >> >>> On 1/16/2008 at 11:44 AM, in message
>> >> >> <d4c...@ma...>, =
"Bernard Li"
>> >> >> <be...@va...> wrote:
>> >> >> > Hi Brad:
>> >> >> >
>> >> >> > On 1/15/08, Brad Nicholes <BNI...@no...> wrote:
>> >> >> >
>> >> >> >>   works for me too.  Check it in :)
>> >> >> >
>> >> >> > Done -- BTW, when I try to run the script in the command prompt =
(i.e.
>> >> >> > python tcpconn.py), it doesn't exit even though I hit ctrl-c -- =
bug or
>> >> >> > feature?
>> >> >> >
>> >> >>
>> >> >> Not sure why it wouldn't exit on ctrl-c but the test portion of =
the script=20
>> > by=20
>> >> > design, goes into an infinite loop.  I guess the while loop could =
check=20
> for=20
>> >=20
>> >> > an exit command, but then it is only for testing purposes anyway =
so=20
>> > whatever=20
>> >> > the mechanism is to kill the script is probably OK.
>> >> >>
>> >> >> Brad
>> >> >>
>> >> >>
>> >>=20
>> >>=20
>> >>=20
>> >>=20
>>=20
>>=20
>>=20
>>=20