Getting (Bad protocol "tcp") errors under Solaris

Getting (Bad protocol "tcp") errors under Solaris

am 12.12.2005 19:59:28 von mfreeman

--nextPart5592645.ZYMlcV9hof
Content-Type: text/plain;
charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

I've got a small script that uses LWP::UserAgent that all of a sudden, or s=
o=20
it seems, just stopped working. It simply takes a file and tries to upload =
it=20
to a CGI. It works under Linux and some of my other Solaris boxes, but this=
=20
one particular machine gives me this error:

dnmcimincw13:web# ./foo.pl moo.zip
CONTENT: 500 (Internal Server Error) Can't connect to rt.innocent.com:443 (=
Bad=20
protocol 'tcp')
Content-Type: text/plain
Client-Date: Mon, 12 Dec 2005 18:48:54 GMT
Client-Warning: Internal response

500 Can't connect to rt.innocent.com:443 (Bad protocol 'tcp')

dnmcimincw13:web#=20


Any ideas?

=2D-=20
Michael J. Freeman
NMS Development
Netco Government Services, Inc.
=2D-
A lack of leadership is no substitute for inaction.


--nextPart5592645.ZYMlcV9hof
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQBDnciZUv/7ffX2I84RAoFFAJ9zhopQPqDbKTVfUhwDHRxfrZ9z1wCe K8cF
PRv6DcvLMofRWaSsL2Ntuhw=
=vAoh
-----END PGP SIGNATURE-----

--nextPart5592645.ZYMlcV9hof--

RE: Getting (Bad protocol "tcp") errors under Solaris

am 12.12.2005 22:39:16 von Matthew.van.Eerde

Michael J. Freeman wrote:
> I've got a small script that uses LWP::UserAgent that all of a
> sudden, or so it seems, just stopped working. It simply takes a file
> and tries to upload it to a CGI. It works under Linux and some of my
> other Solaris boxes, but this one particular machine gives me this
> error:=20
>=20
> dnmcimincw13:web# ./foo.pl moo.zip
> CONTENT: 500 (Internal Server Error) Can't connect to
> rt.innocent.com:443 (Bad protocol 'tcp')
> Content-Type: text/plain

Hmmm...

Well, 443 is the default port for https traffic. Are you specifying
"https" by name anywhere in the script?

--=20
Matthew.van.Eerde (at) hbinc.com 805.964.4554 x902
Hispanic Business Inc./HireDiversity.com Software Engineer

Re: Getting (Bad protocol "tcp") errors under Solaris

am 12.12.2005 23:16:13 von gisle

"Michael J. Freeman" writes:

> 500 Can't connect to rt.innocent.com:443 (Bad protocol 'tcp')

This means that getprotobyname("tcp") fails. On many Unix systems
this is looked up from /etc/protcols. Try:

$ perl -le 'print scalar(getprotobyname("tcp"))'

This ought to print 6 on most systems. If it doesn't try to run it
under truss to figure out what is happening.

--Gisle

Re: Getting (Bad protocol "tcp") errors under Solaris

am 12.12.2005 23:19:46 von mfreeman

--nextPart2101304.uBY3BXZu35
Content-Type: text/plain;
charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

Thank you. I ended up figuring it out using truss. The machine in question =
was=20
recently added as an LDAP client and it was looking for the protocols stuff=
=20
from the directory, which didn't have the data loaded.

On Monday 12 December 2005 16:16, Gisle Aas wrote:
> "Michael J. Freeman" writes:
> > 500 Can't connect to rt.innocent.com:443 (Bad protocol 'tcp')
>
> This means that getprotobyname("tcp") fails. On many Unix systems
> this is looked up from /etc/protcols. Try:
>
> $ perl -le 'print scalar(getprotobyname("tcp"))'
>
> This ought to print 6 on most systems. If it doesn't try to run it
> under truss to figure out what is happening.
>
> --Gisle

=2D-=20
Michael J. Freeman
NMS Development
Netco Government Services, Inc.
=2D-
The world is full of willing people, some willing to work, the rest willing=
to=20
let them.
-- Robert Frost


--nextPart2101304.uBY3BXZu35
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQBDnfeDUv/7ffX2I84RAkChAJ4i0v76HOtPFsmvjr5FOBgYredsKQCd F1xR
aF50SZwZlUnerNuG6b36EiA=
=gZ6A
-----END PGP SIGNATURE-----

--nextPart2101304.uBY3BXZu35--

Re: Getting (Bad protocol "tcp") errors under Solaris

am 13.12.2005 00:20:37 von Andy

On Mon, Dec 12, 2005 at 01:39:16PM -0800, Matthew.van.Eerde@hbinc.com (Matthew.van.Eerde@hbinc.com) wrote:
> Well, 443 is the default port for https traffic. Are you specifying
> "https" by name anywhere in the script?
>

And if you are, do you have Crypt::SSLeay installed? Mech can't do
https without it.

xoox,
Andy

--
Andy Lester => andy@petdance.com => www.petdance.com => AIM:petdance