Overhaul of SlaveLogs
#518
Draft
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
jglick
added a commit
to jglick/support-core-plugin
that referenced
this pull request
Mar 1, 2024
jglick
added a commit
that referenced
this pull request
Mar 18, 2024
* agent-logs sketch * `SlaveLaunchLogsTest.onlineInboundAgent` * `SlaveLaunchLogsTest.offlineAgent` * Javadoc * Exploring `SlaveLaunchLogs` behavior * More `SlaveLaunchLogsTest` * `SlaveLaunchLogsTest.passwords` * Worked out a better `SlaveLaunchLogs`, but depends on patch to `SlaveComputer` * Need to flush logs also for `deletedAgent` * Reverting changes extracted to #518 * Setting a timestamp, switching category * RC * No need to assert that `SupportTestUtils.invokeComponentToString` is non-null * `Security2186Test` failure caused by renamed bundle entry * File handle leak caught by Windows tests * Better handling of rotated logs * More robust way to wait for `Connection terminated` message * `SlaveLaunchLogsTest.offlineAgent` still flaky on Windows * jenkinsci/jenkins#9009 released * Working around lack of JENKINS-72799 to avoid requiring a weekly core * SpotBugs --------- Co-authored-by: Allan Burdajewicz <[email protected]>
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Outdated suggestions cannot be applied.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.
The collection of Java logging from agents is rather antiquated given the prevalence of
Clouds and is missing a lot of critical diagnostics. (CloudBees-internal reference: BEE-46766) Some draft observations:SlaveLogsonly lists information from currently online agents. It might be more useful to run something likeJenkinsRule.RemoteLogDumperto stream agent JUL messages back to a file on the controller, so we could collect these at any time for support bundles, including agents currently defined but disconnected, or agents which have since been deleted. Most agent logs are brief (9aeca36 notwithstanding) so this does not seem unreasonable. Could still collectremoting.logfrom the agent when-workDiris defined, as this would capture messages about connection attempts before the agent is fully online.SlaveLogs.addAgentJulLogRecordsdoes not seem to work well. It stores the logs on the agent, making it impossible to retrieve them if the agent is not currently online, begging the question of why this method even exists (sinceSlaveLogsalso usesComputer.getLogRecordswhich has the same information). Also the logs are stored in the root path, rather than using the workdir when one is defined.Even for an agent currently connected it is not that helpful since it relies on
SupportPlugin.LogInitializercapturing JUL records from the agent JVM afterComputerListener.onOnline, whereasremoting/logs/remoting.log.0inside the workdir includes all the details of the connection logic starting fromhudson.remoting.jnlp.Main.createEngineup throughhudson.remoting.jnlp.Main$CuiListener.statussayingConnected.SmartLogCleaneris deliberately deleting information about historical agents.I would expect custom loggers at fine levels defined on the controller to be honored inside the agent JVM as well, as
log-clidoes.NodeRemoteDirectoryComponentwould includeremoting.log.0but this is not exactly apparent, and only available for one agent at a time by selecting Agent Support.(Extracted from #517 for clarity.)