Re: trouble using sfsagent in MacOS 10.4 Tiger

Joe Cooley <[email protected]> Mon, 19 Dec 2005 16:30:32 -0500
Newsgroups gmane.comp.file-systems.sfs.general
Message-ID <[email protected]>
------=_Part_29890_28767206.1135027832269
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

Hello.

This is a follow-up to Kevin's email.  Has anyone figured out the EOF
problem?  I recently installed the latest sfs from cvs (as of 2005-12-18) o=
n
two different powerbooks, one running Mac OS X 10.4.2 and the other 10.3.9.
The 10.3.9 powerbook is setup as the server and the other is the client.  I
receive the same EOF as mentioned by Kevin when logging in:

>sfskey login [email protected]
sfsagent: EOF from sfscd
sfsagent: exiting
Passphrase for [email protected]/1024: ...

Thanks,
Joe

On 11/1/05, Kevin Fu <[email protected]> wrote:
>
> Sorry, no dice.  The "arg marshaling" problem does not appear.
> Compiling with no optimization whatsoever (env DEBUG=3D-g ./configure)
> results in the same problem:
>
> # /usr/local/sbin/sfscd -d
> sfscd: version 0.8pre, pid 21877
> sfscd: not dropping privileges for debugging
> nfsmounter: version 0.8pre, pid 21880
> nfsmounter: mounted /sfs
> nfsmounter: mounted /sfs/.mnt/wait
> sfsrwcd: version 0.8pre, pid 21878
> sfsrocd: version 0.8pre, pid 21879
>
>
> $ sfsagent [email protected]
> sfsagent: EOF from sfscd
> sfsagent: exiting
>
> Maybe I'll have the joy of running into that problem once I figure
> out this EOF of sfscd.  My SFS client and server are using code from
> CVS several months apart --- so it could be something as simple as a
> protocol mismatch.  Will keep looking.
>
> Cheers,
> -Kevin
>
> On Nov 1, 2005, at 1:53 PM, Jaakko Niemi wrote:
>
> > On Thu, 27 Oct 2005, Kevin Fu wrote:
> >> sfsagent: EOF from sfscd
> >
> >  Hmm, anything in syslog? What if you start sfscd with -d ?
> >  If you see something like:
> >
> >  sfscd: arg marshaling failed (prog 344438, vers 1, proc 2)
> >
> >  you're most propably running into gcc optimization bug that
> >  happens on ppc. Reproduced on at least Linux and couple BSD's.
> >  Easy to work around, just compile with -O0 or -O1 instead of
> >  -O2. Plenty free beers for whoever manages to decent testcase
> >  so that proper bugreport can be made.
> >
> >  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=3D183871
> >
> >                       --j
>
>

------=_Part_29890_28767206.1135027832269
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

Hello.<br><br>This is a follow-up to Kevin's email.&nbsp; Has anyone figure=
d out the EOF problem?&nbsp; I recently installed the latest sfs from cvs (=
as of 2005-12-18) on two different powerbooks, one running Mac OS X 10.4.2 =
and the other=20
10.3.9.&nbsp; The 10.3.9 powerbook is setup as the server and the other is =
the client.&nbsp; I receive the same EOF as mentioned by Kevin when logging=
 in:<br><br>&gt;sfskey login <a href=3D"mailto:[email protected]">[email protected]</a>=
<br>sfsagent: EOF from sfscd
<br>sfsagent: exiting<br>Passphrase for [email protected]/1024: ...<br><br>Thanks=
,<br>Joe<br><br><div><span class=3D"gmail_quote">On 11/1/05, <b class=3D"gm=
ail_sendername">Kevin Fu</b> &lt;<a href=3D"mailto:[email protected]">ke=
[email protected]
</a>&gt; wrote:</span><blockquote class=3D"gmail_quote" style=3D"border-lef=
t: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1=
ex;">Sorry, no dice.&nbsp;&nbsp;The &quot;arg marshaling&quot; problem does=
 not appear.
<br>Compiling with no optimization whatsoever (env DEBUG=3D-g ./configure)<=
br>results in the same problem:<br><br># /usr/local/sbin/sfscd -d<br>sfscd:=
 version 0.8pre, pid 21877<br>sfscd: not dropping privileges for debugging
<br>nfsmounter: version 0.8pre, pid 21880<br>nfsmounter: mounted /sfs<br>nf=
smounter: mounted /sfs/.mnt/wait<br>sfsrwcd: version 0.8pre, pid 21878<br>s=
fsrocd: version 0.8pre, pid 21879<br><br><br>$ sfsagent <a href=3D"mailto:f=
[email protected]">
[email protected]</a><br>sfsagent: EOF from sfscd<br>sfsagent: exitin=
g<br><br>Maybe I'll have the joy of running into that problem once I figure=
<br>out this EOF of sfscd.&nbsp;&nbsp;My SFS client and server are using co=
de from
<br>CVS several months apart --- so it could be something as simple as a<br=
>protocol mismatch.&nbsp;&nbsp;Will keep looking.<br><br>Cheers,<br>-Kevin<=
br><br>On Nov 1, 2005, at 1:53 PM, Jaakko Niemi wrote:<br><br>&gt; On Thu, =
27 Oct 2005, Kevin Fu wrote:
<br>&gt;&gt; sfsagent: EOF from sfscd<br>&gt;<br>&gt;&nbsp;&nbsp;Hmm, anyth=
ing in syslog? What if you start sfscd with -d ?<br>&gt;&nbsp;&nbsp;If you =
see something like:<br>&gt;<br>&gt;&nbsp;&nbsp;sfscd: arg marshaling failed=
 (prog 344438, vers 1, proc 2)
<br>&gt;<br>&gt;&nbsp;&nbsp;you're most propably running into gcc optimizat=
ion bug that<br>&gt;&nbsp;&nbsp;happens on ppc. Reproduced on at least Linu=
x and couple BSD's.<br>&gt;&nbsp;&nbsp;Easy to work around, just compile wi=
th -O0 or -O1 instead of<br>
&gt;&nbsp;&nbsp;-O2. Plenty free beers for whoever manages to decent testca=
se<br>&gt;&nbsp;&nbsp;so that proper bugreport can be made.<br>&gt;<br>&gt;=
&nbsp;&nbsp;<a href=3D"http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=3D1=
83871">http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=3D183871
</a><br>&gt;<br>&gt;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&=
nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbs=
p; --j<br><br></blockquote></div><br>

------=_Part_29890_28767206.1135027832269--