Conversation
Send the haproxy container a HUP instead of restarting it.
rebuild.sh
Outdated
| @@ -0,0 +1,6 @@ | |||
| #!/bin/bash | |||
There was a problem hiding this comment.
Sorry, didn't mean to commit that file.
| if err := p.client.ContainerKill(context.Background(), cnt.ID, "HUP"); err != nil { | ||
| log().Errorf("error reloading container: id=%s err=%s", cnt.ID[:12], err) | ||
| continue | ||
| } |
There was a problem hiding this comment.
Are you sure this catches all cases? A default with log probably wouldn't hurt.
| // update the proxy container status | ||
| cInfo, err := p.client.ContainerInspect(context.Background(), cnt.ID) | ||
| if err != nil { | ||
| log().Errorf("unable to inspect proxy container: %s", err) |
There was a problem hiding this comment.
This will fall through without a continue.
There was a problem hiding this comment.
Ah I didn't see it from the diff. 👍
| } | ||
| case "running": | ||
| log().Debugf("reloading proxy container: id=%s", cnt.ID) | ||
| if err := p.client.ContainerKill(context.Background(), cnt.ID, "HUP"); err != nil { |
There was a problem hiding this comment.
Are you sure HAProxy handles this properly? I believe we've tried HUP before and found that it didn't always reload the config.
There was a problem hiding this comment.
That being said, if we can test it I would love to get rid of the TCP hacks. :)
There was a problem hiding this comment.
Pretty sure it is stable in v1.7.x. I have been using it (outside of interlock) for a while.
| log().Infof("restarted proxy container: id=%s name=%s", cnt.ID[:12], cnt.Names[0]) | ||
| } | ||
|
|
||
| if err := p.resumeSYN(); err != nil { |
There was a problem hiding this comment.
This is needed to safely reload (see above comment).
test/integration: Remove test.hosts fixture
2 participants
Send the haproxy container a HUP instead of restarting it.