On 02/09/2016 01:21 PM, Kyle Mestery wrote: > On Tue, Feb 9, 2016 at 12:18 PM, Russell Bryant <russ...@ovn.org> wrote: >> On 02/09/2016 01:12 PM, Kyle Mestery wrote: >>> On Tue, Feb 9, 2016 at 5:24 AM, Michael <michael.arna...@mymoneyex.com> >>> wrote: >>>> Options for running ovsdb separately is done and also described in the >>>> help file. >>>> >>>> NEWS file is updated. >>>> >>>> I just have to rise a few points: >>>> >>>> - the way thing are running with start/stop daemon from ova-lib require >>>> the daemon to have a pidfile named daemon.pid and this is impossible >>>> to use in our case because we will end up having conflict between >>>> nb and sb db because they both will appear ad ovsdb-server. >>>> >>> I think that needs to be refactored to accept a pid file name so we >>> can have it generate separate pidfiles for the different DBs. >>> >>>> - the way SSL is handles in ovs-ctl make usage of >>>> db:Open_vSwitch,SSL,private_key and I suppose we don’t have this vars >>>> using ovsdb-server without cons.db (please correct me if I’m wrong) >>>> >>> I think fundamentally we need the starting of the NB and SB DBs in >>> ovn-ctl to be done by calling start_ovsdb() from ovs-ctl. See comments >>> below. >> >> That was my suggestion, yes, but the point about about the SSL options >> is a good one. Those are only relevant if the db is the Open_vSwitch db. >> > We could factor those out if the DB is not conf.db easily. My only > concern is duplicating code which essentially is starting ovsdb-server > with different DBs. Keeping it in one place seems reasonable to me. > >> Reading start_ovsdb() in ovs-ctl again, at this point there really isn't >> much more to share so I suppose I retract the suggestion. >> >> One thing ovs-ctl is doing that we don't do in ovn-ctl is check to see >> if daemons are already running. That'd be a helpful enhancement, but is >> not related to this patch. >> > We'd get this for free if we just used a single set of functions to > operate with ovsdb-server. ;)
Sure, but we might be actually increasing the amount of code overall since I'm not sure there's much code to share. I'm not against it if something gets worked out that seems useful. I'm not going to -1 this patch over it at this point, though, myself. -- Russell Bryant _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev