<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
</head>
<body bgcolor="#ffffff" text="#000000">
We are running into several issues configuring a Solaris10 CUPS client
talking to 2 Solaris10 CUPS servers (print1 & print2) in an HA
configuration (i.e. with the same 300 printers defined on each system).<br>
<br>
The sequence of events is as follows:<br>
<br>
1) CUPS is started on print1. The GUI correctly shows all 300 printers.<br>
<br>
2) CUPS is started on print2. The GUI shows 600 printers, including the
300 defined on print1. Also, print1 now shows 600 printers, including
300 from print2. So far all is good.<br>
<br>
3) CUPS is started on client1. The GUI variously shows either 300
printers (what I expect to see) or 600 printers.<br>
If it shows 300 printers, the Device URI for all printers is variously
one of the following:<br>
<b>Device URI:</b>
ipp://print[1/2].xxx.com:631/printers/<printer_name><br>
<b>Device URI:</b> <a class="moz-txt-link-freetext"
href="file:///dev/null">file:///dev/null</a><br>
<font color="#ff0000">Why is this the case?</font><br>
<br>
If it shows 600 printers, the same printer appears twice with the
following Device URIs:<b><br>
Device URI:</b> socket://<printer_name>.xxx.com<br>
<b>Device URI:</b>
ipp://print2.xxx.com:631/printers/<printer_name><br>
<font color="#ff0000">Why aren't the printers recognized as being the
same an collated in some instances?</font><br>
<font color="#ff0000">Why wouldn't the first Device URI show
ipp://print1.xxx.com:631/printers/<printer_name>?</font><br>
<br>
4) CUPS is stopped on one of the servers (print1 or print2). The GUI
shows 300 printers on the remaining server. However, client1 variously
has either all the 300 printers correctly defined, or it may have
anything under 300, or 0 printers. Frequently also, the printers on
client1 will just disappear over time. Snoops & trusses show that
the cups process on the server are correctly broadcasting the printer
info, and the client is correctly receiving them, but somehow ignoring
them. Logs do not give any other info. Restarting the CUPS process on
either the server or the client fixes the problem.<br>
<font color="#ff0000">Why is client1 unable to reconfigure itself with
the broadcast information received from the server that is up?<br>
Why would printers on the client randomly disappear over time, even
though the client and server appear to be communicating?<br>
Why is this fixed automagically as soon as the CUPS process is
restarted?<br>
</font><br>
<br>
This is a new set up - all the machines are on the same subnet and are
not being used for anything else.<br>
<br>
Relevant info in the cupsd.conf file on Servers<br>
----------------------------------------------<br>
<tt>LogLevel info<br>
SystemGroup sys root<br>
Listen localhost:631<br>
Listen print1.xxx.com:631<br>
Listen /print/cups/var/run/cups/cups.sock<br>
# Show shared printers on the local network.<br>
Browsing On<br>
BrowseAddress 192.9.200.255<br>
BrowseOrder allow,deny<br>
BrowseAllow all<br>
BrowseInterval 60<br>
BrowseTimeout 300<br>
ImplicitClasses On<br>
Printcap</tt><br>
<br>
Relevant info in the cupsd.conf file on Client<br>
---------------------------------------------<br>
<tt>LogLevel info<br>
SystemGroup sys root<br>
Listen localhost:631<br>
Listen client1.xxx.com:631<br>
Listen /print/cups/var/run/cups/cups.sock<br>
ImplicitClasses On<br>
Printcap<br>
<br>
Thanks,<br>
-Eli<br>
</tt>
<pre class="moz-signature" cols="72">--
Eli Kleinman
B&H Photo Video, Inc.
420 9TH Avenue
New York, NY 10001 USA
Phone: 212-239-7500 Ext.2154
Email: <a class="moz-txt-link-abbreviated" href="mailto:elik@bhphoto.com">elik@bhphoto.com</a>
</pre>
</body>
</html>