start velia-health after sysrepo is prepared and alarms daemon running
We have seen such errors on live machines with velia-health outputting:
Jun 15 11:25:37 ROADM-LINE-SDN-ID210515 health[32006]: Registered systemd unit watcher for 'dev-ttyq0.device'
Jun 15 11:25:37 ROADM-LINE-SDN-ID210515 health[32006]: Systemd unit 'dev-ttyq0.device' changed state (active plugged)
Jun 15 11:25:37 ROADM-LINE-SDN-ID210515 main[32006]: Fatal error in main: sysrepo::ErrorWithCode
Jun 15 11:25:37 ROADM-LINE-SDN-ID210515 main[32006]: Couldn't send RPC: SR_ERR_UNSUPPORTED
Jun 15 11:25:37 ROADM-LINE-SDN-ID210515 veliad-health[32006]: terminate called after throwing an instance of 'sysrepo::ErrorWithCode'
Jun 15 11:25:37 ROADM-LINE-SDN-ID210515 veliad-health[32006]: what(): Couldn't send RPC: SR_ERR_UNSUPPORTED
... and indeed, velia-health did not have dependency on
sysrepo-ietf-alarmsd so the correct startup order was not guaranteed.
Change-Id: Ia3a8dc3b668fdc82a64b51a467105e5b0f6a5407
diff --git a/package/velia/velia-health.service b/package/velia/velia-health.service
index 61b58c1..d755099 100644
--- a/package/velia/velia-health.service
+++ b/package/velia/velia-health.service
@@ -1,7 +1,7 @@
[Unit]
Description=Health tracking of the whole system
-Requires=multi-user.target
-After=multi-user.target
+Requires=multi-user.target cfg-restore-sysrepo.service czechlight-migrate.service nacm-restore.service sysrepo-ietf-alarms.service
+After=multi-user.target cfg-restore-sysrepo.service czechlight-migrate.service nacm-restore.service sysrepo-ietf-alarms.service
ConditionKernelCommandLine=czechlight
[Service]