I by myself have never tested cups-browsed with that many printers or otherwise very noisy networks. It is also the first time that someone reports that cups-browsed takes a lot of CPU.
cups-browsed is running an event loop and if an event (= broadcast signal from local avahi-daemon or from remote CUPS server) happens, a callback function to create, update, or remove a print queue is executed. Important to know is also what kind of broadcasting the remote servers are doing. Are they running newer CUPS versions and so using Bonjour for broadcasting (cups-browsed is then triggered by the local avahi-daemon) or are these older servers using CUPS broadcasting (cups-browsed uses the legacy CUPS browsing facility then)? Can you deactivate CUPS browsing via cups-browsed.conf? If some printers go away on your local system then you indded have servers using the old method. Does the system load caused by cups-browsed go down now? Perhaps upgrading old servers to have all broadcasting being Bonjour could help. Is there a lot of Bonjour broadcasting in your network, also from other devices, like file servers or so? You can see this running the "avahi-discover" command. Is your local machine sharing CUPS printers? Did you turn on the legacy CUPS broadcasting of your local cups-browsed to share to old clients? Try to turn this off and in case of success, update the CUPS on these clients. Till -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org