LPD printing FROM uw7 TO os5 problem (repost)

[Pat Welch]

The original post has disappeared from my local news server, with no
reply so far. The problem still exists, so:

We just put in our 1st NSC (Non-Stop Cluster) install. UW 7.1.1+IP with
all NSC recommended patches.

We set up remote printing on both boxes, added each of the systems to
the hosts.equiv file and also tried adding .rhosts to the root and
/usr/spool/lp dirs (.rhosts were given 600 mode) on the OS 5.0.5 box.
And we also tried adding the UW7 node to the hosts.lpd file in the
course of trying to make this work.

Printing from the UW7 box to a printer attached to a Netgear print
server box on the UW7 network using LPD is working fine.

But no matter what we try, the OS5 box NAK's any attempt to send print
jobs to it's spooled printers (all attached to Specialix serial ports).
The logs on the UW7 side show a connection then an immediate disconnect
from the OS5 box.

The OS5 side never shows any info in it's log files.

We tried using rcp to copy a text file, with the same lack of success.
rcp uses almost the same type of authentication mechanism for remote
file copies I believe.

We have not tried printing from UW7 to a OS5 attached printer before, so
this could be purely a UW7 issue with OS5 or, as we suspect, NSC is
causing security issues with the OS5 box due to the fact that there are
3 IP addresses underlying the cluster (it's a 2 node Cross Mirror
install).

We tried adding the underlying node names to the hosts.equiv and .rhosts
files as well as the cluster node itself. We also tried the fully
specified hostnames for the cluster and it's nodes.

Pinging and telneting in both directions work with no problem.

Suggestions would be most welcome at this stage of tearing my (limited)
amount of hair out ;-)

--
---------------------------------------------
Pat Welch, UBB Computer Services
SCO Authorized Reseller
Unix/Hardware/BB Sales & Support
(209) 745-1401 Fax: (209) 745-5640
Nationwide pager: (800) 608-7122
E-mail: pat...@inreach.com
----------------------------------------------

[Pat Welch]

Re a suggestion from Roberto Nunin in the UW NG, I added the underlying
node IP addresses to the 'regular' /etc/hosts file.

That did the trick. It does look like the NSC version is not completely
encapsulating the individual IP addresses during secure operations like
remote LPD.

That would appear to be a bug, since NSC is supposed to present a single
cluster IP image to the 'outside' world.