| Summary: | -current behaves differently than 4.X w.r.t rsh from security/krb5 port | ||
|---|---|---|---|
| Product: | Ports & Packages | Reporter: | Tillman Hodgson <tillman> |
| Component: | Individual Port(s) | Assignee: | Cy Schubert <cy> |
| Status: | Closed FIXED | ||
| Severity: | Affects Only Me | ||
| Priority: | Normal | ||
| Version: | Latest | ||
| Hardware: | Any | ||
| OS: | Any | ||
|
Description
Tillman Hodgson
2005-03-02 19:50:02 UTC
Responsible Changed From-To: freebsd-ports-bugs->cy Over to maintainer Followup: I've confirmed that the Heimdal port doesn't exhibit this behaviour -- Kerberos rsh/rshd works fine there. Note that this is the Heimdal port I tested, the Heimdal in the base OS doesn't have the Kerberos bits in rsh. I've been tracking this since December. No fix yet though. I have noticed that the server attempts to connect to a port on the client that is not (or no longer) open. Cheers, Cy Schubert <Cy.Schubert@komquats.com> Web: http://www.komquats.com and http://www.bcbodybuilder.com FreeBSD UNIX: <cy@FreeBSD.org> Web: http://www.FreeBSD.org BC Government: <Cy.Schubert@gems8.gov.bc.ca> "Lift long enough and I believe arrogance is replaced by humility and fear by courage and selfishness by generosity and rudeness by compassion and caring." -- Dave Draper On Fri, Mar 11, 2005 at 12:19:01PM -0800, Cy Schubert wrote: > I've been tracking this since December. No fix yet though. I have noticed > that the server attempts to connect to a port on the client that is not (or > no longer) open. Thanks for tracking this, Cy. My workaround in the meantime has been to install MIT krb5 to /usr/local/krb5 and install the Heimdal port to /usr/local ... and I'm using Heimdal solely for rshd. I'd like to be able to jettison one of the redundant ports. Playing with Heimdal in conjunction with MIT and trying all the combinations gives the following results .,. FreeBSD -current Heimdal port client to FreeBSD 4.x MIT kshd: $ /usr/local/bin/rsh athena uptime 2:47PM up 82 days, 22:38, 4 users, load averages: 0.00, 0.00, 0.05 FreeBSD -current Heimdal port client to FreeBSD 4.x MIT kshd (with encryption): $ /usr/local/bin/rsh -x athena uptime rsh: decrypting data: Decrypt integrity check failed FreeBSD -current Heimdal port client to FreeBSD -current Heimdal port rshd: $ /usr/local/bin/rsh thoth uptime 2:53PM up 9 days, 3:22, 3 users, load averages: 0.03, 0.07, 0.07 FreeBSD -current Heimdal port client to FreeBSD -current Heimdal port rshd (with encryption): $ /usr/local/bin/rsh -x thoth uptime 2:53PM up 9 days, 3:23, 3 users, load averages: 0.01, 0.06, 0.07 FreeBSD 4.x MIT port client to FreeBSD 4.x MIT kshd: [root@athena ~]# rsh athena uptime 2:48PM up 82 days, 22:39, 4 users, load averages: 0.00, 0.00, 0.04 FreeBSD 4.x MIT port client to FreeBSD 4.x MIT kshd (with encryption): [root@athena ~]# rsh athena -x uptime This rsh session is encrypting input/output data transmissions. 2:50PM up 82 days, 22:42, 4 users, load averages: 0.01, 0.00, 0.03 FreeBSD 5.x MIT port client to FreeBSD 4.x MIT kshd: $ /usr/local/krb5/bin/rsh athena uptime socket: protocol error or closed connection in circuit setup FreeBSD 5.x MIT port client to FreeBSD 4.x MIT kshd (using encryption): $ /usr/local/krb5/bin/rsh -x athena uptime socket: protocol error or closed connection in circuit setup So aside from the MIT on 5.x -> anything rsh problems, there also appears to be an encrypted rsh problem (`rsh -x`) between Heimdal and MIT (regardless of FreeBSD version). Since i only have 1 FreeBSD 4.x host, I've just been ignoring that problem. I took some tcpdumps of working a Heimdal on 5.x to Heimdal on 5.x rsh connection and a non-working MIT connection. If they'd be helpful, you can find them at http://www.seekingfire.com/projects/kerberos/tmp/tcpdumps.html. I can also make some ktrace information available if you'd find that useful. -T -- ``These days, thinking someone is out to get you isn't paranoia, it's reality. Lots of people are out to get you. Paranoia is thinking they're all conspiring together.'' -- Unknown State Changed From-To: open->closed As of the latest 5-STABLE and 6-CURRENT, this is no longer a problem. Closing. |