Hide Forgot
Description of problem: cups-browsed in RHEL 7.3 is very old and lacks many features such as high-availability and fail-over. Version-Release number of selected component (if applicable): cups-filters-1.0.35-21.el7.x86_64 How reproducible: N/A Steps to Reproduce: N/A Actual results: N/A Expected results: N/A Additional info: High-availability and fail-over are key features of CUPS browsing. This currently does not work with cups-browsed and RHEL 6 CUPS/1.4.2. I've verified that the upstream (Fedora 26) version of cups-browsed does high-availability and fail-over very nicely with RHEL 6.
Thanks! I'll test it as soon as I can. I'm doing some RHEL 8 cups-browsed HA testing, so it should dovetail nicely into that work. I'll post results as soon as I have them.
I'm sorry for the delay in testing. I finished some of the RHEL 8 tests, and I'll start my RHEL 7 testing today. I've already got a couple of RHEL 8 systems set up, so it shouldn't be much work to set up a RHEL 7 client and see how it interoperates with the RHEL 8 systems.
I can confirm that adding: BrowsePoll rhel8b-master1 BrowsePoll rhel8b-master2 with the following cups-filters package installed # rpm -q cups-filters cups-filters-1.0.35-23.el7.x86_64 does indeed result in the creation an implicit class when the hosts rhel8b-master{1,2} have print queues with the same name: # lpstat -t scheduler is running no system default destination device for raw-socket: implicitclass:raw-socket device for rdn: implicitclass:rdn raw-socket accepting requests since Fri 15 Mar 2019 03:26:03 PM PDT rdn accepting requests since Fri 15 Mar 2019 03:22:41 PM PDT printer raw-socket is idle. enabled since Fri 15 Mar 2019 03:26:03 PM PDT printer rdn is idle. enabled since Fri 15 Mar 2019 03:22:41 PM PDT For what it's worth, I also had to add: CreateRemoteRawPrinterQueues Yes because the queues on rhel8b-master{1,2} are raw queues. I also tested to make sure that the print jobs did go through to the two master print servers as expected. Next I need to test class creation without BrowsePoll.
If I run "cupsdisable raw-socket" on rhel8b-master1, all jobs are routed to rhel8b-master2, as expected.
I can also confirm that setting BrowseRemoteProtocols CUPS on the RHEL 7 client while BrowseProtocols CUPS is set on the master servers results in the queues being found on the client: [root@rhel76-client cups-filters]# lpstat -t scheduler is running no system default destination device for raw-devnull: implicitclass:raw-devnull device for raw-socket: implicitclass:raw-socket device for rdn: implicitclass:rdn raw-devnull accepting requests since Sun 17 Mar 2019 11:51:39 AM PDT raw-socket accepting requests since Sun 17 Mar 2019 11:51:39 AM PDT rdn accepting requests since Sun 17 Mar 2019 11:51:39 AM PDT printer raw-devnull is idle. enabled since Sun 17 Mar 2019 11:51:39 AM PDT printer raw-socket is idle. enabled since Sun 17 Mar 2019 11:51:39 AM PDT printer rdn is idle. enabled since Sun 17 Mar 2019 11:51:39 AM PDT Interestingly, all queues are created as implicitclass queues, even when there's only one printer in the class (raw-devnull only exists on one master print server). That's a little different from the implicit class behavior in RHEL 6 (CUPS 1.4.2), but it is consistent with the behavior or cups-browsed in RHEL 8, so I assume that behavior is correct.
(In reply to Bryan Mason from comment #26) > I can also confirm that setting > > BrowseRemoteProtocols CUPS > > on the RHEL 7 client while > > BrowseProtocols CUPS > > is set on the master servers results in the queues being found on the client: > > [root@rhel76-client cups-filters]# lpstat -t > scheduler is running > no system default destination > device for raw-devnull: implicitclass:raw-devnull > device for raw-socket: implicitclass:raw-socket > device for rdn: implicitclass:rdn > raw-devnull accepting requests since Sun 17 Mar 2019 11:51:39 AM PDT > raw-socket accepting requests since Sun 17 Mar 2019 11:51:39 AM PDT > rdn accepting requests since Sun 17 Mar 2019 11:51:39 AM PDT > printer raw-devnull is idle. enabled since Sun 17 Mar 2019 11:51:39 AM PDT > printer raw-socket is idle. enabled since Sun 17 Mar 2019 11:51:39 AM PDT > printer rdn is idle. enabled since Sun 17 Mar 2019 11:51:39 AM PDT Thank you for all the testing, Bryan! > > Interestingly, all queues are created as implicitclass queues, even when > there's only one printer in the class (raw-devnull only exists on one master > print server). That's a little different from the implicit class behavior > in RHEL 6 (CUPS 1.4.2), but it is consistent with the behavior or > cups-browsed in RHEL 8, so I assume that behavior is correct. cups-browsed started to mark all remote cups queues as implicitclass in auto-discovery since 1.12.0. It uses IPP backend only for real IPP printers (discovered by avahi+IPP2.0 min. requirement+PWG, Apple of PDF supported document format).
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHEA-2019:2220