This repository has been archived by the owner on Aug 1, 2022. It is now read-only.
Proxy deadlock results in CI failing #2133
Comments
41 tasks
geigerzaehler
changed the title
CI failing after proxy goes away
Proxy deadlock results in CI failing
Jul 20, 2021
After some more investigation it looks like the proxy is deadlocking. This only seems to happen if the proxy is reloaded multiple times. Reloading the proxy means that the |
geigerzaehler
added a commit
to radicle-dev/radicle-link
that referenced
this issue
Jul 20, 2021
We ensure that the cleanup function returned by `protocol::accept` from `librad` is called when the `daemon::Peer::run()` future is dropped. This seems to fix deadlocks that occur when the daemon is torn down multiple times. See radicle-dev/radicle-upstream#2133 Signed-off-by: Thomas Scholtes <geigerzaehler@axiom.fm>
This seems to happen also when running Upstream in dev mode via
|
geigerzaehler
added a commit
to radicle-dev/radicle-link
that referenced
this issue
Jul 20, 2021
We ensure that the cleanup function returned by `protocol::accept` from `librad` is called when the `daemon::Peer::run()` future is dropped. This seems to fix deadlocks that occur when the daemon is torn down multiple times. See radicle-dev/radicle-upstream#2133 Signed-off-by: Thomas Scholtes <geigerzaehler@axiom.fm>
geigerzaehler
added a commit
to radicle-dev/radicle-link
that referenced
this issue
Jul 21, 2021
`Peer::run` now properly cleans up the network stack when run to completion. To this end it accepts a shutdown signal as an argument. This seems to fix deadlocks that occur when the daemon is torn down multiple times [1]. Instead of copying the approach from `seed/src/lib.rs` we use a shutdown future and port this approach to the seed. The reason for this is that the seed also did not handle shutdown properly. If a termination signal was sent it did not break the main loop. Moreover we did not react to shutdown requests while awaiting `peer.bind()`. [1]: radicle-dev/radicle-upstream#2133 Signed-off-by: Thomas Scholtes <geigerzaehler@axiom.fm>
geigerzaehler
added a commit
that referenced
this issue
Jul 21, 2021
Update radicle-link to include radicle-dev/radicle-link#731. Fixes #2133 Signed-off-by: Thomas Scholtes <geigerzaehler@axiom.fm>
2 tasks
geigerzaehler
added a commit
that referenced
this issue
Jul 21, 2021
Update radicle-link to include radicle-dev/radicle-link#731. Fixes #2133 Signed-off-by: Thomas Scholtes <geigerzaehler@axiom.fm>
geigerzaehler
added a commit
that referenced
this issue
Jul 21, 2021
Update radicle-link to include radicle-dev/radicle-link#731. Fixes #2133 Signed-off-by: Thomas Scholtes <geigerzaehler@axiom.fm>
geigerzaehler
added a commit
to radicle-dev/radicle-link
that referenced
this issue
Jul 27, 2021
`Peer::run` now properly cleans up the network stack when run to completion. To this end it accepts a shutdown signal as an argument. This seems to fix deadlocks that occur when the daemon is torn down multiple times [1]. Instead of copying the approach from `seed/src/lib.rs` we use a shutdown future and port this approach to the seed. The reason for this is that the seed also did not handle shutdown properly. If a termination signal was sent it did not break the main loop. Moreover we did not react to shutdown requests while awaiting `peer.bind()`. [1]: radicle-dev/radicle-upstream#2133 Signed-off-by: Thomas Scholtes <geigerzaehler@axiom.fm>
geigerzaehler
added a commit
that referenced
this issue
Jul 27, 2021
Update radicle-link to include radicle-dev/radicle-link#731. Fixes #2133 Signed-off-by: Thomas Scholtes <geigerzaehler@axiom.fm>
geigerzaehler
added a commit
that referenced
this issue
Jul 27, 2021
Update radicle-link to include radicle-dev/radicle-link#731. Fixes #2133 Signed-off-by: Thomas Scholtes <geigerzaehler@axiom.fm>
geigerzaehler
added a commit
that referenced
this issue
Jul 27, 2021
Update radicle-link to include radicle-dev/radicle-link#731. Fixes #2133 Signed-off-by: Thomas Scholtes <geigerzaehler@axiom.fm>
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Since the radicle-link upgrade (#2028) our CI has started failing, both on Github and Buildkite.
It seems that this is caused by
radicle-proxy
going not binding to the local API anymore. This causes the tests to error with “Failed to fetch”. This error always occurs after the proxy is reset and a new keystore is created. The last log lines from the proxy areAfter this all subsequent tests fail with “Failed to fetch”.
The text was updated successfully, but these errors were encountered: