On 28.10.21 19:28, Waldemar Brodkorb wrote: > Hi, > > we use z/VM's Live Guest Relocation for our Debian/GNU Linux > servers. Since following commit this is broken: > > d4560150cb47daba50b0d080550ca4bbe80cf3c3 is the first bad commit > commit d4560150cb47daba50b0d080550ca4bbe80cf3c3 > Author: Julian Wiedmann <jwi@xxxxxxxxxxxxx> > Date: Thu Feb 28 18:59:39 2019 +0100 > > s390/qeth: call dev_close() during recovery > > > With Debian/buster and Linux 4.18.x this works fine. > With Debian/bullseye and Linux 5.10.x this does not work anymore. > The reason for that is that all configured routes are lost after > relocating the guest. > This looks to me like a major regression. > > The network is configured via /etc/network/interfaces and executing > systemctl restart networking fixes the issue. This can be done via a > udev rule, but this looks like a rather incomplete workaround to us. > Can you show an example of how your /etc/network/interfaces configures the routes? A quick search [1] points to using 'post-up' statements, is that what you're using or have you tried that already? [1] https://unix.stackexchange.com/questions/321687/what-is-the-best-way-to-add-a-permanent-route > All manually added routing information will be lost anyway. > > And I might not imagine what happens to any firewall connection > tables or ipvs connection tracking information in case of a Live > Guest Relocation. > > So is there any kernel level solution for this you can think of? > As discussed off-list, a plain "ip link set dev eth0 down" gives you the same result. Hence I would recommend to improve your configuration, so that the needed routes are restored when the interface comes up again. > Thanks for any advice and comments, > > best regards > Waldemar >
