You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Nov 25, 2019. It is now read-only.
When an upgrade is deployed, where zero-downtime is active and the supervisor startsecs are changed in the update, the instances are restarted in parallel, but should be restarted serially (since zero-downtime active).
Suspicion:
The script does a bin/supervisorctl update instance1, while instance1 is stopped. When there is no change, nothing happens. But when there is a change in the supervisor config, the new config is loaded an the autostart-setting will cause the instance to be started.
The problem is that supervisor now starts the instance but this is not blocking and the next start instance1 command will be instantly terminated since the instance is already starting. Normally start instance1 would block for startsecs time, so that the instances are properly restarted serially.
This probably needs to be fixed in supervisor.
The text was updated successfully, but these errors were encountered:
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
When an upgrade is deployed, where zero-downtime is active and the supervisor startsecs are changed in the update, the instances are restarted in parallel, but should be restarted serially (since zero-downtime active).
Suspicion:
The script does a
bin/supervisorctl update instance1
, whileinstance1
is stopped. When there is no change, nothing happens. But when there is a change in the supervisor config, the new config is loaded an the autostart-setting will cause the instance to be started.The problem is that supervisor now starts the instance but this is not blocking and the next
start instance1
command will be instantly terminated since the instance is already starting. Normallystart instance1
would block forstartsecs
time, so that the instances are properly restarted serially.This probably needs to be fixed in supervisor.
The text was updated successfully, but these errors were encountered: