[OpenAFS-devel] Tracing NameToId
Derrick Brashear
shadow@gmail.com
Mon, 6 Apr 2009 13:23:08 -0400
--000e0cd40204eaa06e0466e626e9
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
On Mon, Apr 6, 2009 at 1:16 PM, Barry Fawthrop <barry@isscp.com> wrote:
> Derrick Brashear wrote:
> > On Mon, Apr 6, 2009 at 12:59 PM, Barry Fawthrop <barry@isscp.com> wrote:
> >
> >> Okay so a "thread timer" is called and another thread is called.
> >>
> >> I have tried sending the debug information but moderator rejected due to
> >> size
> >>
> >
> > sure. you put it in afs or the web and give us a url or a path, and it's
> not
> > huge. problem solved.
> >
> >
> >> So I'm going try again to use personally
> >>
> >>
> > My Server IP is 24.227.64.44 now where near the 182.179.15.100
> >
> > what's the client's IP? is nat in play?
> >
> >
> >
> >
>
> Yes Nat is in play My client is 192.168.0.11
>
> Wireshark on the client can see the packets being sent to the server
> Kerberos and TGS
> and the initial NameToId Request to the Server
>
> Wireshark sees multiple responses back from the server
> but the client does not see them it must have already crashed.
>
> What more besides the trace given in the last email is required ?
>
i have no idea. i don't have time to decode the assembly and see what's up,
and really, it's not worth spending a whole lot of time on lwp anyway; in
the nearish future openafs will just use pthreads. if you want to do
experimental stuff, that might be possible now: define USE_PTHREADS when
building liblwp (e.g. in src/lwp) and then use that liblwp. or, better yet,
stop linking liblwp, librx, librxkad and company and link (only) libafsrpc,
libafsauthent, libafsutil and libpthread.
--
Derrick
--000e0cd40204eaa06e0466e626e9
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
<br><br><div class=3D"gmail_quote">On Mon, Apr 6, 2009 at 1:16 PM, Barry Fa=
wthrop <span dir=3D"ltr"><<a href=3D"mailto:barry@isscp.com">barry@isscp=
.com</a>></span> wrote:<br><blockquote class=3D"gmail_quote" style=3D"bo=
rder-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding=
-left: 1ex;">
<div><div></div><div class=3D"h5">Derrick Brashear wrote:<br>
> On Mon, Apr 6, 2009 at 12:59 PM, Barry Fawthrop <<a href=3D"mailto:=
barry@isscp.com">barry@isscp.com</a>> wrote:<br>
><br>
>> Okay so a "thread timer" is called and another thread is=
called.<br>
>><br>
>> I have tried sending the debug information but moderator rejected =
due to<br>
>> size<br>
>><br>
><br>
> sure. you put it in afs or the web and give us a url or a path, and it=
's not<br>
> huge. problem solved.<br>
><br>
><br>
>> So I'm going try again to use personally<br>
>><br>
>><br>
> My Server IP is 24.227.64.44 =A0now where near the 182.179.15.100<br>
><br>
> what's the client's IP? is nat in play?<br>
><br>
><br>
><br>
><br>
<br>
</div></div>Yes Nat is in play =A0My client is 192.168.0.11<br>
<br>
Wireshark on the client can see the packets being sent to the server<br>
Kerberos and TGS<br>
and the initial NameToId Request to the Server<br>
<br>
Wireshark sees multiple responses back from the server<br>
but the client does not see them it must have already crashed.<br>
<br>
What more besides the trace given in the last email is required ?<br>
</blockquote><div><br>i have no idea. i don't have time to decode the a=
ssembly and see what's up, and really, it's not worth spending a wh=
ole lot of time on lwp anyway; in the nearish future openafs will just use =
pthreads. if you want to do experimental stuff, that might be possible now:=
define USE_PTHREADS when building liblwp (e.g. in src/lwp) and then use th=
at liblwp. or, better yet, stop linking liblwp, librx, librxkad and company=
and link (only) libafsrpc, libafsauthent, libafsutil and libpthread.<br>
<br></div></div>-- <br>Derrick<br>
--000e0cd40204eaa06e0466e626e9--