[JBoss JIRA] (WFLY-11391) server.log is erased when application fails to start
by Brian Stansberry (Jira)
[ https://issues.jboss.org/browse/WFLY-11391?page=com.atlassian.jira.plugin... ]
Brian Stansberry reassigned WFLY-11391:
---------------------------------------
Component/s: Logging
Assignee: (was: Jason Greene)
This needs clear reproducer instructions that don't require figuring out how to install keycloak. Simply putting an unresolvable expression in some config attribute in our stock standalone.xml doesn't cause this, at least not with current master.
> server.log is erased when application fails to start
> ----------------------------------------------------
>
> Key: WFLY-11391
> URL: https://issues.jboss.org/browse/WFLY-11391
> Project: WildFly
> Issue Type: Bug
> Components: Logging
> Affects Versions: 14.0.1.Final, 13.0.0.Final
> Reporter: Vlastislav Ramik
> Priority: Major
>
> After upgrade from Wildfly 11 to 13 we face an issue described in KEYCLOAK-8821.
> For us it is not a blocker but could be for wildfly, so for that case creating this one. I'm not 100% sure the cause is in the wildfly itself, I need to investigate it further before I could tell more.
--
This message was sent by Atlassian Jira
(v7.12.1#712002)
7 years, 7 months
[JBoss JIRA] (LOGMGR-86) Create an exception filter
by David Lloyd (Jira)
[ https://issues.jboss.org/browse/LOGMGR-86?page=com.atlassian.jira.plugin.... ]
David Lloyd commented on LOGMGR-86:
-----------------------------------
Of course this can be easily done with a custom Filter implementation.
> Create an exception filter
> --------------------------
>
> Key: LOGMGR-86
> URL: https://issues.jboss.org/browse/LOGMGR-86
> Project: JBoss Log Manager
> Issue Type: Feature Request
> Reporter: James Perkins
> Priority: Major
>
> For now this is just a thought so that users can filter messages based on an exception in the stack trace. Currently it's not possible to use a regex filter as the stack trace is not used in when testing for matches and it probably shouldn't be used.
--
This message was sent by Atlassian Jira
(v7.12.1#712002)
7 years, 7 months
[JBoss JIRA] (WFLY-10864) MP Health reports UP when there is port collision for port 8080
by Brian Stansberry (Jira)
[ https://issues.jboss.org/browse/WFLY-10864?page=com.atlassian.jira.plugin... ]
Brian Stansberry commented on WFLY-10864:
-----------------------------------------
IMHO if we do anything like this we should keep it very simple, a la our OS health probes at https://github.com/jboss-openshift/cct_module/blob/master/os-eap-probes/a..., which basically check server-state, whether there were errors at boot, whether any deployments are failed. I could also see tracking whether MSC stability could not be reached when executing a mgmt op and failing if that had occurred, as the server is in an unknowable state at that point.
If the server booted cleanly and it hasn't subsequently gotten into an MSC unstable state, then if something isn't working "right" that either means a a valid configuration, but just not what the user really wanted, or the user executed a management op that failed but they said the flag that prevented rollback. The latter I see little point in trying to account for as it's a silly thing to do on a system that someone would want to monitor for health.
Checking for valid but not really what was wanted configuration (oops, I forgot to add an undertow listener on 808!!) seems like endless work.
Note I believe the checks I linked above would find the port collision issue as it would prevent clean boot.
> MP Health reports UP when there is port collision for port 8080
> ---------------------------------------------------------------
>
> Key: WFLY-10864
> URL: https://issues.jboss.org/browse/WFLY-10864
> Project: WildFly
> Issue Type: Bug
> Components: MP Health
> Reporter: Rostislav Svoboda
> Assignee: Jeff Mesnil
> Priority: Critical
>
> MP Health reports UP when there is port collision for port 8080.
> I have simple python app to acquire port 8080 and once it runs I start WF server
> {code}
> import SimpleHTTPServer
> import SocketServer
> PORT = 8080
> Handler = SimpleHTTPServer.SimpleHTTPRequestHandler
> httpd = SocketServer.TCPServer(("127.0.0.1", PORT), Handler)
> print "serving at port", PORT
> httpd.serve_forever()
> {code}
> WF reports Address already in use /127.0.0.1:8080, deployments are not deployed.
> Accessing http://localhost:9990/health/ reports "outcome":"UP"
> I think this is wrong and DOWN should be reported.
> Spec is misleading because it says {{A producer without health check procedures installed MUST returns positive overall outcome (i.e. HTTP 200)}} but it silently assumes server started correctly.
> I believe authors of the spec expected that if there is some trouble like port collision the server doesn't start, WildFly is a bit further as it allows to boot even if some services are not started properly.
--
This message was sent by Atlassian Jira
(v7.12.1#712002)
7 years, 7 months