Hi konstantin, There is a non-default re-sync functionality available in the openflow plugin. If you use this, you should be in a position to restart the controller, or a compute in any order (at least from a netvirt and related features perspective) Tunnel deletes are not well handled today, I,e if you remove a switch, and delete a tep in the controller, the system goes out of sync, we are adding an ovsdb re-sync feature in fluorine mainly for this usecase, but if you are not changing config, it should not block you currently. I would suggest you enable the resync, and report any issues found to the group.
Thanks, daya -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Konstantin Agouros Sent: Wednesday, April 25, 2018 2:05 AM To: odl netvirt dev <[email protected]> Subject: [netvirt-dev] Proper way to shutdown and restart everything? Hi, I am under the impression that the whole system is not resilient to a proper shutdown/restart of the whole ensemble. So is the proper way to first shutdown the compute nodes and then ODL or the other way around? Also: Should the tunnels-ports on the compute nodes be left as be or deleted before starting again? At the moment I get an inconsistent state, where VMs on some compute nodes have network and others not. Thanks, Konstantin -- Dipl-Inf. Konstantin Agouros aka Elwood Blues. Internet: [email protected] Altersheimerstr. 1, 81545 Muenchen, Germany. Tel +49 89 69370185 ---------------------------------------------------------------------------- "There is no 'dead' in team!" - Sameen Shaw _______________________________________________ netvirt-dev mailing list [email protected] https://lists.opendaylight.org/mailman/listinfo/netvirt-dev _______________________________________________ openflowplugin-dev mailing list [email protected] https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev
