Description of problem: From debian-based distributions (Ubuntu in my case) I can't connect to none-debian based distros with remote tunel over SSH (qemu+ssh) - virsh / virt-manager. I attach patch from Ubuntu (Debian) which is applied on nc. This patch adds q option [quit after EOF on stdin and delay of secs (-1 to not quit)]. How reproducible: Ubuntu 9.10 (10.04) as administrator desktop system. RedHat as server system. Steps to Reproduce: 1. From Ubuntu to RedHat (xxx.xxx.xxx.xxx) $ virsh -c qemu+ssh://root.xxx.xxx/system Actual results: Can't connect to libvirt via virsh / virt-manager from Ubuntu. Expected results: Can connect to libvirt via virsh / virt-manager from Ubuntu. Additional info: I use Fedora and CentOS on my host machines (servers) and Ubuntu on my desktop. Reported as bug: for Ubuntu: https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/517478 for CentOS: http://bugs.centos.org/view.php?id=4188
I'm setting proper flags for product management to consider inclusion in RHEL 5.6
Created attachment 389086 [details] nc -q option patch
Created attachment 399858 [details] autodetect nc params This patch was written by Marc Deslauriers from Ubuntu and inspired by http://hg.fedorahosted.org/hg/virt-manager/rev/1f781890ea4a and I hope that Marc sent this patch to upstream.
I don't really like the idea of adding more options not accepted upstream. Moreover, EL5+ and Fedora netcats quit after SIGHUP so I don't really see a problem there. If anything, libvirt shouldn't rely on such made-up option. Could you elaborate on what's actually expected?
Apparently Debian patched their client. There's no need to add an exception to server for this; closing.