-
Notifications
You must be signed in to change notification settings - Fork 42
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
xud doesn't shut down gracefully with active rpc streaming calls #699
Comments
Can you provide more details/steps to reproduce? If I start up xud, connect to peers, and then run |
Yes. Will try to generate a log
…On Fri, 23 Nov 2018 at 7:59 Daniel McNally ***@***.***> wrote:
Can you provide more details/steps to reproduce? If I start up xud,
connect to peers, and then run xucli shutdown it does end the process. Do
you see different?
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#699 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AJQ0cibYu-a7y0nSAFsF86Y5Sm28CV9Hks5ux464gaJpZM4Ytibn>
.
|
Thanks, I do see at least one thing that could keep us from ending the process (the peer reconnection timers) so that might be the issue here. |
This is what I find in the log: 11/23/2018, 6:05:52 AM [GLOBAL] info: XUD is shutting down Easy to reproduce on test1.xud |
You need to have a peer connected to the node to see the problem. |
Also note that it is not working also with master. |
This properly cancels the timer to retry a connection to a peer when the peer is closed. Not canceling the timer could interfere with gracefully shutting down `xud`, as the process would continue waiting for the timer after. Fixes #699.
This properly cancels the timer to retry a connection to a peer when the peer is closed. Not canceling the timer could interfere with gracefully shutting down `xud`, as the process would continue waiting for the timer. Fixes #699.
Can you retry your test with the branch from #701? |
Not solved. Suggest you start by reproducing the problem. |
Ok I'll have to revisit tomorrow. It's not as simple as having a peer connected as my first try to reproduce did not have this issue even with peers connected. Can you share the full log including all peers you connected to? Are you just starting xud and then shutting down, or doing anything else in between? |
Log file:
|
@sangaman The problem is due to a stream connection which is open to the server. You can reproduce by the following:
The process will not exit until you kill |
Thanks for figuring that out! I was stuck since I'd never been calling |
This tracks all active streaming gRPC calls and closes them when the gRPC server is shut down. Previously, active streaming calls would prevent `xud` from shutting down gracefully. Fixes #699.
Will check thebPr later today.
Looks like this is not the only reason why shutdown may not work. I face it
on a different configuration even without the stream. Will try to
understand why and let you know.
…On Fri, 30 Nov 2018 at 9:24 Daniel McNally ***@***.***> wrote:
Thanks for figuring that out! I was stuck since I'd never been calling
streamorders - that makes sense. I'll open a PR for a fix.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#699 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AJQ0ct-MNfUo0XXTb8f7W6ehpru_gwpUks5u0N0SgaJpZM4Ytibn>
.
|
Yes. |
This tracks all active streaming gRPC calls and closes them when the gRPC server is shut down. Previously, active streaming calls would prevent `xud` from shutting down gracefully. Fixes #699.
xucli shutdown does not bring the process down and does not close the active peer connections.
The text was updated successfully, but these errors were encountered: