On Tue, Apr 02, 2019 at 12:27:47PM +0200, Christian Ehrhardt wrote: > This is related, but not the same as [1] and about the new defaults in > regard to QoS. > Due to upstream change [2] the new defaults now mean that ssh is > broken if connecting to (or through) a system running on VMWare. > > More details can be found on the Ubuntu bug [3] where this was > reported initially. > As [1] I'd ask for a revert of that change until iptables (for [1]), > vmware (for this bug) and others if identified later will be fixed > (the actual issue is in those components, so sooner or later this > should no more be needed to be reverted). > > Final FYI - Fedora [4] took the decision to force the issue by > sticking with the new default. > After all the "workaround" is user controllable being: > > Configure your client to use the old defaults permanently in > => /etc/ssh/ssh_config > Host * > IPQoS lowdelay throughput > # You might want to limit to your VMware based systems > > Or per command via: > $ ssh IPQoS="latency throughput" user@host
If it were just the VMware issue, then my inclination would be to leave OpenSSH as it is: it's proprietary software and the only leverage we have to get them to fix it is to have their customers complaining. However, the iptables issue in #923879 seems thornier and it's outside my field of expertise. I'm slightly leaning towards reverting this on a temporary basis for buster, but CCing debian-devel: does anyone have opinions on this? > I thought it is better to file it to make you aware of this extra > context to [1] so that you can consider those two requests to revert > [2] together. > > [1]: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=923879 > [2]: > https://anongit.mindrot.org/openssh.git/commit/?id=5ee8448ad7c306f05a9f56769f95336a8269f379 > [3]: https://bugs.launchpad.net/ubuntu/+source/openssh/+bug/1822370 > [4]: https://bugzilla.redhat.com/show_bug.cgi?id=1624437#c8 Thanks, -- Colin Watson [cjwat...@debian.org]
