ruediger-maass commented on a change in pull request #4299: Avoid that actions
do not stop after action timeout when logging heavily
URL:
https://github.com/apache/incubator-openwhisk/pull/4299#discussion_r260728858
##########
File path:
common/scala/src/main/scala/org/apache/openwhisk/core/containerpool/logging/DockerToActivationFileLogStore.scala
##########
@@ -123,7 +123,11 @@ class DockerToActivationFileLogStore(system: ActorSystem,
destinationDirectory:
container: Container,
action: ExecutableWhiskAction):
Future[ActivationLogs] = {
- val logs = container.logs(action.limits.logs.asMegaBytes,
action.exec.sentinelledLogs)(transid)
+ // wait for a sentinel only if no container (developer) error occurred to
avoid
+ // that log collection continues if the action code still logs after
timeout
+ val sentinel = action.exec.sentinelledLogs &&
!activation.response.isContainerError
Review comment:
We have changed the implementation to use an activation annotation `timeout`
(boolean) to indicate action timeouts as proposed by @sven-lange-last.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
