agent: shutdown if the http server goroutine exits #8234
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
dnephin
force-pushed
the
dnephin/shutdown-on-http-server-error
branch
from
a77b864
to
97ff981
Compare
dnephin
force-pushed
the
dnephin/shutdown-on-http-server-error
branch
from
97ff981
to
f41b3e8
Compare
dnephin
commented
Jul 2, 2020
| start := func(proto string, addrs []net.Addr) error { | ||
| listeners, err := a.startListeners(addrs) | ||
| if err != nil { | ||
| return err | ||
| } | ||
| ln = append(ln, listeners...) |
There was a problem hiding this comment.
This was moved up from line 1178. I believe the old code would not properly close the listener that caused the error because it was not appended to the list until after an error is returned.
| }() | ||
|
|
||
| select { | ||
| case addr := <-notif: |
There was a problem hiding this comment.
This select/case no longer exists. All this was doing was waiting for a goroutine to start. The server would not actually be started when the addr is sent on the channel.
With this change, a failure to server will cause a shutdown, so we now properly handle the case without the need for the notif channel.
dnephin
force-pushed
the
dnephin/shutdown-on-http-server-error
branch
from
f41b3e8
to
1ac74a4
Compare
dnephin
force-pushed
the
dnephin/unembed-HTTPServer-Server
branch
from
9a3617e
to
df40882
Compare
dnephin
force-pushed
the
dnephin/shutdown-on-http-server-error
branch
from
1ac74a4
to
0a7a2c8
Compare
hanshasselberg
approved these changes
Aug 14, 2020
Remove Server field from HTTPServer. The field is no longer used.
dnephin
force-pushed
the
dnephin/shutdown-on-http-server-error
branch
from
0a7a2c8
to
6ca45e1
Compare
|
I amended the last commit to fix a merge conflict and added more godoc to |
dnephin
added a commit
that referenced
this pull request
Sep 4, 2020
https server. In #8234 I changed a few tests to use TestAgent.HTTPAddr() to find the addr used in the test. Due to the way HTTPAddr() was implemented these tests were passing, but I think the pass was incidental. HTTPAddr() was not matching any servers, and was instead returning the last server, which happened to be the one these tests wanted. This commit fixes the implementation of HTTPAddr to panic if no match was found. The tests which require an HTTPS server are changed to use a new firstAddr() to look up the correct address.
This was referenced Sep 4, 2020
15 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Fixes #6677
x/sync/errgroup is designed to handle exactly this problem. I have used it in the past and found that it handled the problem well.
Eventually we may be able to use a single
errgroup.Groupas the core of the agent. That would require that every goroutine shutdown on a cancelled context (instead of callingShutdown). Until that is possible I've put it into a newapiServersstruct and exposed the channel so thatcommandcan be notified when one of them exits, and proceed with the shutdown.As a follow up to this PR I think we can move at least the
dnsServersintoapiServers, which should further shrinkagent.Agent. Maybe it would be appropriate to move thegrpcServersin as well?The retryJoiner for Lan/Wan pools also works the same way. So using this errgroup for them would help shrink the Agent even further.