| Message ID | 73ce112f3554cae9e3e5ece28e5ffb499ed6a4f6.1614944209.git.tredaelli@redhat.com |
|---|---|
| State | Deferred |
| Headers | show |
| Series | [ovs-dev] rhel: increase timeout of ovsdb-server.service to 300s | expand |
On Fri, Mar 05, 2021 at 12:47:05PM +0100, Timothy Redaelli wrote: > In some scenarios starting ovsdb-server takes more than 90 seconds and > so ovsdb-server.service is marked as failed. > > Signed-off-by: Timothy Redaelli <tredaelli@redhat.com> I'm surprised it takes so long. What leads to that? Is it a bug in ovsdb-server?
On Fri, 5 Mar 2021 10:34:36 -0800 Ben Pfaff <blp@ovn.org> wrote: > On Fri, Mar 05, 2021 at 12:47:05PM +0100, Timothy Redaelli wrote: > > In some scenarios starting ovsdb-server takes more than 90 seconds and > > so ovsdb-server.service is marked as failed. > > > > Signed-off-by: Timothy Redaelli <tredaelli@redhat.com> > > I'm surprised it takes so long. What leads to that? Is it a bug in > ovsdb-server? Hi, The scenario is an openshift (RHOCP4.6) server in low-performance mode after a reboot of the server. After the failure systemd automatically restarts ovsdb-server.service (that starts successfully at this time), but due to [1] openvswitch.service and ovs-vswitchd.service are not started and they needs to be started manually. In the meanwhile systemd implements the RFE in [1], I think changing the timeout to 300 seconds (the same timeout we already use for ovs-vswitchd.service) is a good workaround. [1] https://github.com/systemd/systemd/issues/18856
Hi Timothy, Timothy Redaelli <tredaelli@redhat.com> writes: > In some scenarios starting ovsdb-server takes more than 90 seconds and > so ovsdb-server.service is marked as failed. > > Signed-off-by: Timothy Redaelli <tredaelli@redhat.com> > --- > NOTE: Currently systemd will not start openvswitch.service > and ovs-vswitchd.service if ovsdb-server.service is automatically restarted > after a failure, but this is a feature that needs to be implemented in systemd > itself (systemd#18856). > --- > rhel/usr_lib_systemd_system_ovsdb-server.service | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/rhel/usr_lib_systemd_system_ovsdb-server.service b/rhel/usr_lib_systemd_system_ovsdb-server.service > index 98338b9df..ed6419f31 100644 > --- a/rhel/usr_lib_systemd_system_ovsdb-server.service > +++ b/rhel/usr_lib_systemd_system_ovsdb-server.service > @@ -29,3 +29,4 @@ ExecStop=/usr/share/openvswitch/scripts/ovs-ctl --no-ovs-vswitchd stop > ExecReload=/usr/share/openvswitch/scripts/ovs-ctl --no-ovs-vswitchd \ > ${OVS_USER_OPT} \ > --no-monitor restart $OPTIONS > +TimeoutSec=300 > -- > 2.29.2 waiting for the vswitchd feature, the change LGTM. Acked-by: Paolo Valerio <pvalerio@redhat.com>
diff --git a/rhel/usr_lib_systemd_system_ovsdb-server.service b/rhel/usr_lib_systemd_system_ovsdb-server.service index 98338b9df..ed6419f31 100644 --- a/rhel/usr_lib_systemd_system_ovsdb-server.service +++ b/rhel/usr_lib_systemd_system_ovsdb-server.service @@ -29,3 +29,4 @@ ExecStop=/usr/share/openvswitch/scripts/ovs-ctl --no-ovs-vswitchd stop ExecReload=/usr/share/openvswitch/scripts/ovs-ctl --no-ovs-vswitchd \ ${OVS_USER_OPT} \ --no-monitor restart $OPTIONS +TimeoutSec=300
In some scenarios starting ovsdb-server takes more than 90 seconds and so ovsdb-server.service is marked as failed. Signed-off-by: Timothy Redaelli <tredaelli@redhat.com> --- NOTE: Currently systemd will not start openvswitch.service and ovs-vswitchd.service if ovsdb-server.service is automatically restarted after a failure, but this is a feature that needs to be implemented in systemd itself (systemd#18856). --- rhel/usr_lib_systemd_system_ovsdb-server.service | 1 + 1 file changed, 1 insertion(+)