Tuesday, 2015-12-15

*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has quit IRC (Quit: Leaving...)01:22
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has joined #wildfly-swarm01:34
*** wangc <wangc!~wangc@redhat/jboss/wangc> has joined #wildfly-swarm02:20
*** jclingan <jclingan!~jclingan@redhat/jclingan> has quit IRC (Quit: My Mac has gone to sleep. ZZZzzz…)04:33
*** hbraun <hbraun!~hbraun@pdpc/supporter/professional/hbraun> has joined #wildfly-swarm07:23
jbossbotnew jira [SWARM-203] Track core metrics about default distribution [Backlog (Unresolved) Task, Major, Unassigned] https://issues.jboss.org/browse/SWARM-20307:38
*** hbraun <hbraun!~hbraun@pdpc/supporter/professional/hbraun> has quit IRC (Quit: afk)08:01
*** hbraun <hbraun!~hbraun@pdpc/supporter/professional/hbraun> has joined #wildfly-swarm08:11
*** agoncal <agoncal!~agoncal@179.21.90.92.rev.sfr.net> has joined #wildfly-swarm08:37
*** agoncal <agoncal!~agoncal@179.21.90.92.rev.sfr.net> has quit IRC (Ping timeout: 240 seconds)08:44
*** wangc <wangc!~wangc@redhat/jboss/wangc> has quit IRC (Quit: Leaving)08:55
*** agoncal <agoncal!~agoncal@158.17.90.92.rev.sfr.net> has joined #wildfly-swarm09:07
*** agoncal <agoncal!~agoncal@158.17.90.92.rev.sfr.net> has quit IRC (Ping timeout: 256 seconds)09:15
*** agoncal <agoncal!~agoncal@162.17.90.92.rev.sfr.net> has joined #wildfly-swarm09:29
*** wangc <wangc!~wangc@redhat/jboss/wangc> has joined #wildfly-swarm09:32
*** wangc <wangc!~wangc@redhat/jboss/wangc> has quit IRC (Remote host closed the connection)09:55
*** agoncal <agoncal!~agoncal@162.17.90.92.rev.sfr.net> has quit IRC (Read error: Connection reset by peer)09:59
*** helio-frota <helio-frota!~heliof@8.80.249.186.araujosat.com.br> has joined #wildfly-swarm10:09
*** agoncal <agoncal!~agoncal@176.16.90.92.rev.sfr.net> has joined #wildfly-swarm10:11
helio-frotamorning!10:27
hbraunmorning10:32
helio-frotaI'm going to review each example today10:33
*** ro14nd <ro14nd!~ro14nd@p5DD1D1CF.dip0.t-ipconnect.de> has joined #wildfly-swarm10:35
helio-frotahbraun, no need to create a JIRA for it right ?10:35
hbraunhelio-frota not sure what you mean?10:36
helio-frotato 'review the examples'10:36
hbraunno10:36
helio-frotaok10:36
hbraunwe just need jiras for issues that we don't immediately follow up upon, or that should be documented in case other people encounter the same problem/requirement10:36
hbraunbut what do you actually mean with "review"?10:37
hbraunverify it's working?10:37
helio-frota1) read the readme.md and see if makes sense10:37
ro14ndhi ;-).I have an issue with connecting the Jolokia JVM agent to a swarm app (https://github.com/wildfly-swarm/wildfly-swarm-examples/issues/41).  Guess it was the wrong place to open issue (because it got closed). What is the canonical place for opening such an issue ?10:37
jbossbotgit issue [wildfly-swarm-examples] (closed) rhuss Exception when running with Jolokia JVM agent https://github.com/wildfly-swarm/wildfly-swarm-examples/issues/4110:37
jbossbotgit issue [wildfly-swarm-examples] (closed) rhuss Exception when running with Jolokia JVM agent https://github.com/wildfly-swarm/wildfly-swarm-examples/issues/4110:37
helio-frota2) verify if it's working10:38
helio-frota3) try to fix10:38
hbraunro14nd I closed it because from your comment I thought is is working?10:39
hbraunwhile you can still use github, we'd prefer the issues reported here: https://issues.jboss.org/projects/SWARM/issues10:39
ro14ndok, will do so.10:40
ro14ndno, its not working.10:40
hbraunro14nd should we reopen it and move it over the jira?10:40
hbraunro14nd what are you trying to achieve?10:40
ro14ndJolokia can operate in two mode: Either with -javaagent during startup or dynamically attaching to a running process (via the attach API). Latter works, former not.10:40
hbraunand the motivation for attaching jolokia is?10:41
ro14ndhbraun: yes, we should fix that,. The reason is that the fabric8 base Docker images come with the Jolokia enabled by default (and so currently no swarm will work with a fabric8 base Docker image out of the box).10:41
hbraunah, ok10:42
ro14ndhawt.io and the OpenShift console connect to Java based images via Jolokia.10:42
hbraunro14nd maybe you can it the rationale to the issue description?10:42
ro14ndhbraun: ok, will do it.10:43
hbraunbbrowning_away have you mean aware of the fabric8 base docker image?10:44
ro14ndhbraun: It's not only the fabric8 base images, but also the official jboss-fuse-6/fis-karaf-openshift and jboss-fuse-6/fis-java-openshift images (OSE s2i images in openshift, release during this week ...)10:51
*** agoncal <agoncal!~agoncal@176.16.90.92.rev.sfr.net> has quit IRC (Ping timeout: 240 seconds)10:56
jbossbotnew jira [SWARM-204] Exception when connecting Jolokia JVM agent. [Backlog (Unresolved) Bug, Major, Unassigned] https://issues.jboss.org/browse/SWARM-20411:03
*** mgoldmann|away is now known as mgoldmann11:03
*** maschmid <maschmid!maschmid@nat/redhat/x-rwpxqjyjlkvwxfxl> has joined #wildfly-swarm11:06
hbraunro14nd could be related to the way WF sets up the modules loader11:13
hbraunlet me see if I can find something11:13
hbraunro14nd sounds similar to this one: http://zeroturnaround.com/forums/topic/error-must-set-java-util-logging-manager/11:17
jbossbotTitle: Topic: Error: must set java.util.logging.manager | zeroturnaround.com11:17
helio-frotahbraun, also i have a clojure new example to do a PR later11:18
hbraunhelio-frota cool11:18
hbraunro14nd what happens if explicitly set the -Djava.util.logging.manager=org.jboss.logmanager.LogManager?11:20
hbraunalong with -javaagent ?11:20
hbraunro14nd but reading what other people said, I still think that the javaagent somehow introduces another log manager11:21
ro14ndhbraun: yeah, seems like the jrebel issue is similar.11:22
hbraunif still in doubt this afternoon, we can try to ping dmlloyd11:22
ro14ndhbraun: will try ... and also check for the logging manager (but I'm pretty sure I didn't introduced it :)11:23
hbraunmaybe a transitive dep?11:23
ro14ndhbraun: jolokia does have a single dependency: json-simple (and is 400k). json-simple doesn't have any deps.11:27
ro14ndi mean, jolokia is 400k (including json-simple)11:27
ro14ndget now 'java.lang.ClassNotFoundException: org.jboss.logmanager.LogManager' ...11:28
hbraunro14nd yes, it's reported here: https://issues.jboss.org/browse/WFLY-89511:30
jbossbotjira [WFLY-895] JBoss crashes when run with AspectJ java agent [Resolved (Won't Fix) Bug, Major, David Lloyd] https://issues.jboss.org/browse/WFLY-89511:30
hbraunro14nd i am trying to make sense of the comments as we speak11:30
ro14ndwill do some remote debugging ....11:30
hbraunbut the workaround deals with changing standalone.xml11:30
hbraunsomething that we want to prevent11:30
hbraunro14nd did "-Djava.util.logging.manager=org.jboss.logmanager.LogManager" lead to the classnotfound?11:31
ro14ndyep11:31
ro14ndhttps://gist.github.com/881ff23aec637a0d38a511:32
ro14ndah, the java http sever which I start uses jul ...11:33
ro14ndsun.net.httpserver.ServerImpl.<init>(ServerImpl.java:92)11:33
hbraunro14nd but I assume the agent uses java.util.logging?11:34
ro14ndJolokia use the sun.net.httpserver for exposing metrics (included in the JDK). The http server in turn uses java.util.logging  when starting ...11:35
hbraunriht, it's what stuart said: " This problem will occur with any java agent that uses java.util.logging. This is because the java agent does not have access to jboss-logmanager classes, so even if the java.util.logging.manager property is set the classes will not be available.11:35
hbraun"11:35
ro14ndso, Jolokia doesn't use j.u.l directly, but indirectly,11:36
ro14ndsounds difficult to solve.11:37
hbrauni can talk to dmlloyd about it. maybe we can tweak the properties in swarm interannly11:38
hbrauns/interannly/internaly11:38
hbraunsimilar to the changes that have been to standalone.xml here:  https://issues.jboss.org/browse/WFLY-89511:38
jbossbotjira [WFLY-895] JBoss crashes when run with AspectJ java agent [Resolved (Won't Fix) Bug, Major, David Lloyd] https://issues.jboss.org/browse/WFLY-89511:38
hbrauni.e11:38
hbraunadd JBOSS_MODULES_SYSTEM_PKGS="org.jboss.logmanager" to the beginning of standalone.conf11:38
hbraunadd -Djava.util.logging.manager=org.jboss.logmanager.LogManager to JAVA_OPTS at the end of standalone.conf11:38
hbraunadd -Xbootclasspath/p:$JBOSS_HOME/modules/org/jboss/logmanager/main/jboss-logmanager-1.3.1.Final.jar to JAVA_OPTS at the end of standalone.conf (make sure the JAR name and path matches your installation)11:38
hbraunobviously that's nothing jolokia or fabric8 can or should do11:39
hbraunro14nd but why again does it need the agent? technically? what's the purpose?11:40
hbraundies it expose the REST interface to the JMX server?11:40
ro14ndhbraun: its used to do a deep inspection of Java application in OpenShift and Kubernetes. I.e. the OpenShift management console provides a link 'connect' to a container running a Java application which has Jolokia at port 8778. From that hawt.io is started and provided access to your Camel routes, ActiveMQ queues, JMX Mbeans, ....11:44
ro14ndAll is secured via https and client certs. was quite some work to get it in ;-)11:45
ro14ndThis works perfectly for flat classpath apps like spring boot but also karaf based containers.11:46
ro14ndI thought it would be cool if this works for swarm apps, too.11:46
ro14ndtbh i think its quite dangerous to assume that JVM agents must not use j.u.l before swarm or jboss modules kicks in. Its too much of a restriction.11:48
helio-frotamissing license added to pom.xml files, going todo a PR12:03
*** rbenevides <rbenevides!~rbenevid@redhat/jboss/RafaelBenevides> has joined #wildfly-swarm12:17
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has quit IRC (Remote host closed the connection)12:24
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has joined #wildfly-swarm12:25
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has joined #wildfly-swarm12:26
*** maschmid <maschmid!maschmid@nat/redhat/x-rwpxqjyjlkvwxfxl> has quit IRC (Quit: Ex-Chat)12:28
helio-frotaSWARM-167 was solved :D12:30
helio-frotahttps://issues.jboss.org/browse/SWARM-16712:31
helio-frotai did the test against  datasource-deployment example12:31
helio-frotahas no plugin version but is downloading the right version of the plugin12:32
hbraunro14nd i can see your point. but to be honest, i don't know enough to answer that question. I guess there have been good reasons for that design approach12:33
hbraunthe best way to move forward would be to discuss it with the wildfly guys. in particular dmlloyd and maybe james also12:34
helio-frotafolks I can not close an issue that I created on Jira12:35
helio-frotaSWARM-167 can be closed , to reproduce just run it in datasource-deployment example:12:36
helio-frotamvn help:effective-pom | grep -A2 -B1  "wildfly-swarm-plugin"12:36
helio-frotawill show the right version of wf-swarm-plugin12:36
dmlloydit's not a restriction that we chose12:43
helio-frota?12:47
helio-frotaon Æsh project the jira has 2 more buttons: 'edit' and 'resolve issue'12:48
helio-frotanot a problem about the 'resolve issue' based on organization of the team/rules etc.. but 'edit' is good to add wdyt ?12:49
*** samueltauil <samueltauil!~samueltau@187.57.107.66> has joined #wildfly-swarm12:53
*** maschmid <maschmid!maschmid@nat/redhat/x-uifbpydumhqmsmyp> has joined #wildfly-swarm12:53
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has quit IRC (Remote host closed the connection)13:02
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has joined #wildfly-swarm13:02
hbraunro14nd still here?13:03
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has quit IRC (Ping timeout: 250 seconds)13:07
hbraunro14nd what happens to java process that already include jolokia?13:07
hbraunis that detected somehow?13:08
hbrauni am wondering if the situation changes when we would embed jolokia in swam13:08
hbraunas as last resort: changing jolokia is probably more likely then changing wildfly13:09
ro14ndyes, I'm still there :)13:11
ro14ndembedding Jolokia would work of course, because then Jolokia is started late.13:11
ro14ndhowever we would need to change the base image specifically for wildfly / swarm, too.13:12
ro14ndalso, setting up client cert authentification etc. would need some extra configuration on the wildfly side.13:12
ro14ndbut is Jolokia really the only agent having this problem ? I think of jrebel, jprofiler, prometheus ....13:13
ro14ndIf I could, I would omit the call to j.u.l. But I can't because it is within JDK code.13:14
hbraunyes, that's why I said the last resort13:15
hbraunit would require you to embed and change the http server13:15
ro14ndwhy is changing wildfly hard ?13:15
hbraunwell, compared to jolokia is much bigger and has tons of backwards compatibility constraints13:16
hbraunlogging is quiet central and very exposed component13:16
hbrauneven if we manage to convince the WF team to change it (not sure it's even possible) it would take a _very_ long time to appear in a release13:17
ro14ndhbraun: sorry, jolokia won't change the http server because of this issue. Including e.g. netty or undertow as http server (which btw could refer to j.u.l as well) would more then double the size of the agent.13:17
ro14ndhbraun: ok, understand. is this true for swarm, too ?13:17
ro14ndswarm is still alpha ;-)13:18
hbrauni am not saying it should. just outlining the different perspectives13:18
*** lance|afk is now known as lanceball13:18
hbraunswarm ,is based on wildfly 10, which is naything then alpha13:18
hbraunro14nd and: don't shoot the messenger.13:19
ro14ndhbraun: yeah, i see your point.13:19
ro14ndnp :)13:19
hbraunit's a complex question to answer, i am just trying to explore possible ways to move forward13:19
ro14ndwe probably need to let this bubble up.13:19
ro14ndbtw, its is not critical for me. I just tried something out and wanted to report the issue.13:20
hbraunwell, if you OS3 assume that jolokia can manage java processes this way, then it will be crucial, no?13:21
hbraundon't know how hawt.io and jolokia are positioned in the OS3 ecosystem13:21
hbraunif both are optional then we just consider it to be a limitation13:22
hbraunif both are central to the management of java processes then better look for ways to improve the situation13:22
*** agoncal <agoncal!~agoncal@243.202.154.77.rev.sfr.net> has joined #wildfly-swarm13:22
ro14ndhbraun: yes of course, I really would love to improve that. However, I don't this at the moment what can be done from Jolokia's side (except of exchanging the HTTP server which is not realistic).13:30
ro14ndI'll bring that up in the fabric8 team to rise the awareness.13:30
hbraungood13:30
ro14ndbtw, afaik, the jboss-eap-7-docker use Jolokia in the same way.13:31
ro14ndmgoldmann: how do you include Jolokia in the CE images ? Also as an agent with -javaagent ?13:31
hbraunro14nd guess then there must be some substantial changes in the WF distro of that image13:33
hbraunmaybe something similar as outlined in https://issues.jboss.org/browse/WFLY-89513:34
jbossbotjira [WFLY-895] JBoss crashes when run with AspectJ java agent [Resolved (Won't Fix) Bug, Major, David Lloyd] https://issues.jboss.org/browse/WFLY-89513:34
hbraunro14nd but since we are talking about "java -jar foobar.jar" it think our options are somewhat limited13:35
hbraunin regular wildfly the problem is mitigated by changing the standalone.sh script13:36
hbraunin partcicular the dep on the jbosslogging is passed as a filesystem ref13:36
ro14ndhbraun: but isnt' the point of swarm to exactly allow this: java -jar bla-swarm.jar ? (not talking about wildfly here ,...)13:37
hbraunyes, I did only compare the fix in stock Wf to what we can do in swarm13:38
hbrauni.e this is not possible: -Xbootclasspath/p:$JBOSS_HOME/modules/org/jboss/logmanager/main/jboss-logmanager-1.3.1.Final.jar13:38
jbossbotgit [wildfly-swarm-examples] push master 4387fa6.. helio-frota adding license missing in pom.xml files13:38
jbossbotgit [wildfly-swarm-examples] push master 869e8d7.. Bob McWhirter Merge pull request #42 from helio-frota/ADD-LICENSE...13:38
jbossbotgit [wildfly-swarm-examples] push master URL: http://github.com/wildfly-swarm/wildfly-swarm-examples/compare/c4e2d7b...869e8d713:38
mgoldmannro14nd: yes - javaagent13:38
*** kenfinnigan <kenfinnigan!~ken@redhat/jboss/kenfinnigan> has joined #wildfly-swarm13:38
hbraunmgoldmann did tun into the problems described in https://issues.jboss.org/browse/WFLY-89513:38
jbossbotjira [WFLY-895] JBoss crashes when run with AspectJ java agent [Resolved (Won't Fix) Bug, Major, David Lloyd] https://issues.jboss.org/browse/WFLY-89513:38
hbraun?13:38
projectodd-ciStarting build #46 for job wildfly-swarm-examples-uberjar (previous build: SUCCESS)13:38
mgoldmannhbraun: yep, known issue13:39
hbrauns/tun/you run13:39
hbraunmgoldmann how did you work around it?13:39
mgoldmannhttps://gist.github.com/goldmann/92b3f205c13f9004fdb413:39
mgoldmannupdated, refresh13:40
hbraunthanks, that's what i expected13:40
mgoldmannbasically - a mess13:40
hbraunmgoldmann haha, yes13:41
hbrauna mess13:41
projectodd-ciStarting build #87 for job wildfly-swarm-examples (previous build: SUCCESS)13:41
ro14ndmgoldmann: thanks :)13:42
mgoldmannyw13:42
ro14ndhbraun: this could be a workaround we could live with, too (e,.g. by setting a specific workaround) but I wonder whether it would work with swarm fat jar, too ?13:43
ro14ndSince everything is within the jar file, should the whole jar then be added to the bootclasspath ?13:44
hbraunro14nd no, i don't think it's desireable13:44
hbraunro14nd you can try, but i would expect other errors then13:44
hbraunnot sure if the sawrm jar actually can be used that way13:45
hbraunthe libs included are packaged in subfolders afaikt13:45
ro14ndyeah, especially because its not really a flat jar but contains other jar as well (requiring a dedicared classloader for that013:45
hbraunit's not just the classes repackaged13:45
ro14nddedicated13:45
hbraun+113:45
hbraunyes, that's what I mean13:45
*** agoncal <agoncal!~agoncal@243.202.154.77.rev.sfr.net> has quit IRC (Read error: Connection reset by peer)13:45
hbraunro14nd for the Xclasspath to work we would need an artifact13:46
ro14ndok to summarize: for wildfly we have a workaround, for swarm probably not.13:46
hbraunyes13:47
hbraunbecause in WF you have the artifacts on the filesystem13:47
hbraunso the Xclasspath workaround works13:47
hbraunplus all the other rubbish yu have to add13:48
hbraunro14nd how do the agents actually work?13:48
hbraunare they in the same process?13:48
hbraunro14nd could a delay of the jolokia agent work?13:50
ro14ndyes, they are in the same process13:51
hbrauni.e. defer it after the VM has started?13:51
ro14ndno, a delay won't work because premain() (that the method called by the Java agent API) blocks until finished.13:51
ro14ndI could add an extra thread, though.13:52
hbraunisn't the part in premain simply moved to agentmain?13:52
hbraunsee http://docs.oracle.com/javase/7/docs/api/java/lang/instrument/package-summary.html # Starting Agents After VM Startup13:52
jbossbotTitle: java.lang.instrument (Java Platform SE 7 )13:52
hbraunjust brainstorming, i never implemented an agent myself13:53
projectodd-ciProject wildfly-swarm-examples-uberjar build #46: SUCCESS in 14 min: https://projectodd.ci.cloudbees.com/job/wildfly-swarm-examples-uberjar/46/13:53
projectodd-ci00hf11: adding license missing in pom.xml files13:53
hbraungoes together with the Agent-Class manifest entry13:53
hbraunro14nd for jolokia jmx access this seems to be the more reliable options anyway, no? i mean what happens when the actual java process hasn't full started and you invoke on jolokia?13:55
* hbraun is more generally speaking about java processes managed by jolokia13:56
ro14ndwell, there might be simply some MBean not available. The JMX subsystem is already there.13:56
projectodd-ciProject wildfly-swarm-examples build #87: SUCCESS in 14 min: https://projectodd.ci.cloudbees.com/job/wildfly-swarm-examples/87/13:56
projectodd-ci00hf11: adding license missing in pom.xml files13:56
hbraunbut it wouldn't hurt either, would it?13:56
ro14ndhbraun: is there a singular event which I could wait for before starting the HttpServer ?13:56
ro14ndno, wouldn't and I would add this :-)13:56
ro14ndif we find a good point in time when to start up.13:56
hbraunan event in WF you mean?13:57
hbrauni don't understand the last question13:57
hbraun"a good point in time"?13:57
hbrauni would think that the agent architecture notifies the agent when the process main() has completed. which seems to be the contract for agentmain()13:58
*** bbrowning_away is now known as bbrowning13:59
*** agoncal <agoncal!~agoncal@167.20.90.92.rev.sfr.net> has joined #wildfly-swarm14:04
hbraunro14nd what about the agentmain() lifecycle? would that be an option or does jolokia depend on being initialised in premain()?14:08
bobmcwhbraun: ro14nd: I know it doesn't solve you particular problem, but we have wildfly-swarm-jolokia14:09
bobmcwwhich installs it as a .war14:09
bobmcwand works well with -swarm14:09
hbraunbobmcw jolokia attaches itself to java processes in OS314:10
hbraunin certain (prebuild) images14:10
hbraundocker images14:10
hbraunusing -javagent14:10
bobmcwcan fabric8 not communicate with a jolokia installed a different way?14:10
bobmcwbbrowning's OS3 images don't do -javagent, but could do -swarm, for instance14:10
bbrowningyou can set environment variables w/ my OS3 images to pass JVM and/or -jar options14:11
hbraunbobmcw  afaik it could, but in OS3 the SSL setup is quiet involved14:11
bobmcwah14:11
hbraunto get access to jolokia from within the OS3 management interfaces14:11
hbraunhence they've put it into preconfigured fabric8 and OS3 images. is this correct ro14nd ?14:12
bobmcwif getting agent-based is hard, but you find a manual way to solve it, perhaps we should embody that into wildfly-swarm-container somehow14:12
bobmcwor our plugin14:12
ro14ndhbraun: just being in a meeting. will have a look later ....14:13
hbraunro14nd ok14:13
hbraunttyl14:13
ro14ndhbraun: premain() is for the Agent api, agentmain() for the attach API. Jolokia uses both.14:21
hbraunro14nd when do you start the http server?14:22
hbraunro14nd you have the most knowledge about jolokia ...14:22
ro14ndin both ;-). agentmain() is not called with -javaagent but when you attach dynamically to a process.14:23
ro14ndhttps://github.com/rhuss/jolokia/blob/95f928f084f60ad67738c37ab4368cd82ec2bae9/agent/jvm/src/main/java/org/jolokia/jvmagent/JvmAgent.java#L7614:23
hbraunah, see my knowledge about agent is limited14:23
ro14ndthese are simply two different entry points.14:23
hbrauni did simply refer to the docs http://docs.oracle.com/javase/7/docs/api/java/lang/instrument/package-summary.html14:24
jbossbotTitle: java.lang.instrument (Java Platform SE 7 )14:24
bobmcwtoday I'm going to move the keycloak-server-service into its own repo14:24
bobmcwor into a wildfly-swarm-servers repo14:24
ro14ndwith a 'good point in time' I mean a certain state I can query or so to know when to start up. What I don't want to do is to wait a fixed amount of time for starting up.14:24
*** jclingan <jclingan!~jclingan@redhat/jclingan> has joined #wildfly-swarm14:24
hbraunro14nd know i see what you mean. "starting agents after the vm" implies "connecting" after the VM started?14:25
ro14ndhbraun: right, but agentmain() is called when you attach from the outside of the process.14:25
ro14ndyes14:25
ro14ndso which jolokia you can also attach (and detach) to a running process.14:26
hbraunro14nd ok, tnx for the explanation14:26
ro14nda solution could be to not use -javaagent but wait until the swarm process has started (checkong e.g. some log) and then attach to this process.14:26
ro14ndbut this is quite complicated and I'm afraid also a bit fragile.14:27
hbraunyes, that sounds very brittle14:27
maschmidro14nd, given that on openshift the jolokia agent needs to generate its own RSA keypair on startup which prolongs the statup time significantly, having it boot in parallel to the process itself would probably be a good idea14:28
bobmcwro14nd: if we made some CountDownLatch() available for your await()?14:28
bobmcwSwarm.BOOT_LATCH.await()14:28
bobmcwor something?14:28
ro14ndmaschmid: do you have any number how long it takes longer ? I didn't experienced any delay so far because of key generation ...14:29
bobmcwwe could actually flip that well before the container is fully up, if all we really need is logmanager configured14:29
bobmcwsince we do that in our Container() ctor14:29
maschmidro14nd, depends... on VMs with depleted entropy pools it can take even tens of seconds :)14:29
bobmcwtypically line #1 of a user's main()14:29
ro14ndbobmcw: this could help, however would tie this to swarm (jolokia is mostly a general purpose agent). Jolokia has some means to detect certain servers (so called "ServerDetectors") to adapt to specific behaviour but I'm afraid they are initialized later (after the HttpServier is already up).14:31
bobmcwro14nd: ah, hm.14:31
ro14ndbobmcw: Jolokia is setup before main() (in premain())14:31
bobmcwro14nd: right, but was wondering if it could pause until a latch then counts-down, to deal with its httpserver/logging bits14:32
* bobmcw has no idea, honestly14:32
ro14ndmaschmid: i see, probably I have  a good entropy source here (probably by typing too fast to much nonsense :)14:37
*** hbraun is now known as hbraun_meeting14:42
ro14ndmaybe one could have a dedicate 'log initializer' agent which could be used to do this initialization ? If specified before other agents this would be running first (and then it would be really the first one touching the log system). just brainstorming, too ;-) .....14:51
*** agoncal <agoncal!~agoncal@167.20.90.92.rev.sfr.net> has quit IRC (Read error: Connection reset by peer)14:53
bobmcwhttps://github.com/wildfly-swarm/wildfly-swarm-servers14:54
jbossbotgit [wildfly-swarm] push master 993b917.. Bob McWhirter Remove the keycloak-server-service to https://github.com/wildfly-swarm/wildfly-swarm-servers14:58
jbossbotgit [wildfly-swarm] push master URL: http://github.com/wildfly-swarm/wildfly-swarm/commit/993b9176214:58
projectodd-ciStarting build #398 for job wildfly-swarm (previous build: SUCCESS)14:58
hbraun_meetingbobmcw +1 on https://github.com/wildfly-swarm/wildfly-swarm-servers14:59
hbraun_meetingmaybe something for APIMan too?14:59
ytanabebobmcw, cool. I tested it and I found it fits ro14nd docker-maven-plugin e.g. https://github.com/emag/wildfly-swarm-tour/blob/master/complete/lifelog-docker/keycloak-server/pom.xml#L24-L8714:59
bobmcwhbraun_meeting: sounds good, and hawkular for sure15:00
bobmcwytanabe: hm, maybe we should produce docker images as a normal part of the -servers build15:01
hbraun_meetingbobmcw +115:01
ytanabebobmcw, that's so nice. +115:01
jbossbotnew jira [SWARM-205] expose a way to set additional modules and repos when testing via Arquillian [Backlog (Unresolved) Task, Major, Toby Crawley] https://issues.jboss.org/browse/SWARM-20515:11
bobmcwtcrawley: is the first half of that no jboss-deployment-structure.xml or our *Archive.addModule() in your @Deployment?15:12
tcrawleyno, it's about adding modules/ to the -swarm.jar so that they can be found at runtime15:13
bobmcwhttps://github.com/wildfly-swarm/wildfly-swarm/blob/master/container/api/src/main/java/org/wildfly/swarm/container/JBossDeploymentStructureContainer.java#L3315:13
bobmcwah15:13
tcrawleyvia -Dswarm.build.modules15:14
bobmcwsomething in BuildTool then?15:14
bobmcwtool.addModule( 'whatever.foo.bar', 'main' )15:14
bobmcwincluding transitives?  We have a ModuleAnalyzer that'd help iff so.15:14
tcrawleyBuildTool already has .addAdditionalModules("path/to/modules/dir") - this is about getting the module.xml's into the jar15:15
*** agoncal <agoncal!~agoncal@239.204.154.77.rev.sfr.net> has joined #wildfly-swarm15:16
tcrawleythis is adding a way for the Arq container to know what to pass to .addAdditionalModules, since the <modules> setting of the package mojo isn't available at surefire time15:16
jbossbotgit [wildfly-swarm] push master 6931708.. Bob McWhirter Remove integs/ since wildfly-swarm-examples now has tests.15:18
jbossbotgit [wildfly-swarm] push master URL: http://github.com/wildfly-swarm/wildfly-swarm/commit/6931708b315:18
bobmcwhttps://wildfly-swarm.gitbooks.io/wildfly-swarm-users-guide/content/servers.html15:18
jbossbotTitle: Servers | WildFly Swarm User's Guide15:18
bobmcwah15:19
bobmcwokie dokie15:19
*** bobmcw <bobmcw!~bobmcw@redhat/jboss/bobmcw> has quit IRC ()15:20
*** agoncal <agoncal!~agoncal@239.204.154.77.rev.sfr.net> has quit IRC (Read error: Connection reset by peer)15:22
*** hbraun_meeting <hbraun_meeting!~hbraun@pdpc/supporter/professional/hbraun> has quit IRC (Quit: afk)15:24
*** bobmcw <bobmcw!~bobmcw@va-67-233-86-120.dhcp.embarqhsd.net> has joined #wildfly-swarm15:27
*** bobmcw <bobmcw!~bobmcw@va-67-233-86-120.dhcp.embarqhsd.net> has quit IRC (Changing host)15:27
*** bobmcw <bobmcw!~bobmcw@redhat/jboss/bobmcw> has joined #wildfly-swarm15:27
*** ChanServ sets mode: +o bobmcw15:27
*** agoncal <agoncal!~agoncal@239.204.154.77.rev.sfr.net> has joined #wildfly-swarm15:29
helio-frotabobmcw, great addition -servers +115:36
helio-frotahi folks just reminder: https://issues.jboss.org/browse/SWARM-16715:42
helio-frotathis can be closed ^15:42
helio-frotato check: 1) go to wildfly-swarm-examples/datasource/datasource-deployment15:43
helio-frota2) mvn help:effective-pom | grep -A2 -B1  "wildfly-swarm-plugin"15:43
helio-frotano version of wf-swarm-plugin is defined and getting the current version15:43
dmlloydbobmcw: I almost have the pluggable resolver working.  Introduced a Heisenbug though...15:48
*** jclingan <jclingan!~jclingan@redhat/jclingan> has quit IRC (Quit: My Mac has gone to sleep. ZZZzzz…)15:50
lanceballa heisenbug?15:50
lanceballis that a bug that, when you look at it, it changes state?15:50
dmlloydyes :)15:51
dmlloydbobmcw: https://github.com/dmlloyd/jboss-modules/tree/maven-bobmcw15:51
dmlloydif you could check over those changes and verify that the public APIs work for you I'd appreciate it15:51
*** agoncal <agoncal!~agoncal@239.204.154.77.rev.sfr.net> has quit IRC (Ping timeout: 240 seconds)15:53
*** maschmid <maschmid!maschmid@nat/redhat/x-uifbpydumhqmsmyp> has quit IRC (Quit: Ex-Chat)15:54
*** agoncal <agoncal!~agoncal@148.16.90.92.rev.sfr.net> has joined #wildfly-swarm15:55
bobmcwdmlloyd: will do after lunch more than likely16:10
* bobmcw has meetings until noon, then a sandwich16:10
dmlloydOK.  Once you give your OK I will merge it and the other settings.xml fix and then prepare (I guess) a CR tag.16:11
bobmcwokie dokie16:11
bobmcwyou have ctomc's settings.xml changes? He took that JIRA16:12
*** hbraun <hbraun!~hbraun@pdpc/supporter/professional/hbraun> has joined #wildfly-swarm16:14
*** hbraun <hbraun!~hbraun@pdpc/supporter/professional/hbraun> has quit IRC (Client Quit)16:14
bobmcwI've got to jostle CI some, now for -servers, and adjust -examples to use org.wildfly.swarm.servers:keycloak16:19
*** helio-frota <helio-frota!~heliof@8.80.249.186.araujosat.com.br> has quit IRC (Quit: Leaving)16:20
projectodd-ciStarting build #1 for job wildfly-swarm-servers16:23
bobmcwdmlloyd: can you allow https://github.com/wildfly-swarm/wildfly-swarm-servers to push events to jbossbot please?16:23
dmlloyd%set github/projects/wildfly-swarm/wildfly-swarm-servers/channels #wildfly-swarm16:24
jbossbotchannels set to #wildfly-swarm16:24
dmlloyddone-zo16:24
projectodd-ciProject wildfly-swarm build #398: SUCCESS in 1 hr 26 min: https://projectodd.ci.cloudbees.com/job/wildfly-swarm/398/16:25
projectodd-ciBob McWhirter: Remove the keycloak-server-service to16:25
bobmcwthanks!16:25
projectodd-ciStarting build #399 for job wildfly-swarm (previous build: SUCCESS)16:25
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has joined #wildfly-swarm16:25
*** jclingan <jclingan!~jclingan@redhat/jclingan> has joined #wildfly-swarm16:26
projectodd-ciStarting build #88 for job wildfly-swarm-examples (previous build: SUCCESS)16:26
projectodd-ciStarting build #47 for job wildfly-swarm-examples-uberjar (previous build: SUCCESS)16:26
projectodd-ciProject wildfly-swarm-servers build #1: SUCCESS in 4 min 48 sec: https://projectodd.ci.cloudbees.com/job/wildfly-swarm-servers/1/16:28
bobmcwdmlloyd: MavenResolver looks good... but not yet pluggable in MavenArtifactUtil?16:30
bobmcwor used in MAU.resolveArtifact()?16:30
bobmcwdmlloyd: maybe ArtifactCoords#getRelativePath() also?16:34
bobmcwand https://github.com/dmlloyd/jboss-modules/blob/d489c61b307cc0b68fb8bf767d1c5b0df65debf1/src/main/java/org/jboss/modules/maven/MavenResolver.java#L52-L53 is very specific to a specific resolver16:35
bobmcwreally, entire javadoc on that is specific to default resolver16:35
bobmcwthis is good: https://github.com/dmlloyd/jboss-modules/blob/d489c61b307cc0b68fb8bf767d1c5b0df65debf1/src/main/java/org/jboss/modules/maven/MavenResolver.java#L52-L5316:36
dmlloydah I just copied the javadoc16:36
dmlloydit could probably use some brushing up16:37
dmlloydI kept the MavenArtifactUtil methods as the default mechanism16:37
bobmcwI don't see traversing a List<MavenResolver> but I could be reading too quickly16:37
* bobmcw isn't sure how to configure his fancy resolver into it16:37
dmlloydI don't traverse, but it is trivially possible to create an aggregating resolver16:37
dmlloydbasically MAU doesn't use this - this uses MAU16:38
dmlloydand the XML parser uses this16:38
dmlloydyou use the MAU when you want the default behavior only, and you use MR when you want pluggable behavior (like the XML parser does)16:38
bobmcwa-ha: https://github.com/dmlloyd/jboss-modules/blob/d489c61b307cc0b68fb8bf767d1c5b0df65debf1/src/main/java/org/jboss/modules/xml/ModuleXmlParser.java#L22116:38
dmlloydso you can pass your MR into the XML parser16:38
bobmcwI'd pass ours16:38
dmlloydyeah16:38
bobmcwright on16:39
dmlloydI turned it all inside-out :)16:39
*** agoncal <agoncal!~agoncal@148.16.90.92.rev.sfr.net> has quit IRC (Ping timeout: 256 seconds)16:40
bobmcwwhatever works, works.16:40
dmlloydseems OK then?16:42
bobmcwI think so. If we can drop-in against WF 10.CR2 or whatever, we can give it a try from a SNAPSHOT or whatnot.16:43
bobmcwmodulo heisenbug16:43
dmlloydgreat.16:44
jbossbotgit [wildfly-swarm-examples] push master 9fe4e1a.. Bob McWhirter Move ribbon-secured to use the newly-moved keycloak server.16:55
jbossbotgit [wildfly-swarm-examples] push master URL: http://github.com/wildfly-swarm/wildfly-swarm-examples/commit/9fe4e1aa316:55
*** projectodd-ci <projectodd-ci!~PircBotx@ec2-54-81-247-237.compute-1.amazonaws.com> has quit IRC (Ping timeout: 250 seconds)17:16
*** helio-frota <helio-frota!~heliof@8.80.249.186.araujosat.com.br> has joined #wildfly-swarm17:18
*** samueltauil <samueltauil!~samueltau@187.57.107.66> has quit IRC (Ping timeout: 240 seconds)17:25
*** projectodd-ci <projectodd-ci!~PircBotx@ec2-54-145-107-102.compute-1.amazonaws.com> has joined #wildfly-swarm17:30
*** agoncal <agoncal!~agoncal@212-198-109-24.rev.numericable.fr> has joined #wildfly-swarm17:37
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has quit IRC (Remote host closed the connection)17:45
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has joined #wildfly-swarm17:46
*** bbrowning is now known as bbrowning_away17:49
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has quit IRC (Ping timeout: 240 seconds)17:50
*** lanceball is now known as lance|afk18:20
*** helio-frota <helio-frota!~heliof@8.80.249.186.araujosat.com.br> has quit IRC (Read error: Connection reset by peer)18:46
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has joined #wildfly-swarm18:46
*** bbrowning_away is now known as bbrowning18:50
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has quit IRC (Ping timeout: 260 seconds)18:54
jbossbotgit [wildfly-swarm] push master 98f907b.. Toby Crawley Add sysprop to dump Arquillian swarmjar for debugging19:03
jbossbotgit [wildfly-swarm] push master 82e5815.. Toby Crawley Allow setting modules and repos when arq testing [SWARM-205]19:03
jbossbotgit [wildfly-swarm] push master 1da8f09.. Toby Crawley Include additional modules when resolving build dependencies.19:03
jbossbotgit [wildfly-swarm] push master faebab1.. Toby Crawley Add test to verify additional repos and modules in arq build [SWARM-205]19:03
jbossbotgit [wildfly-swarm] push master a063a52.. Toby Crawley Expose debugging remote arq container via sysprop19:03
jbossbotgit [wildfly-swarm] push master URL: http://github.com/wildfly-swarm/wildfly-swarm/compare/6931708...a063a5219:03
jbossbotjira [SWARM-205] expose a way to set additional modules and repos when testing via Arquillian [In Progress (Unresolved) Task, Major, Toby Crawley] https://issues.jboss.org/browse/SWARM-20519:03
projectodd-ciStarting build #400 for job wildfly-swarm (previous build: SUCCESS)19:04
*** jbott <jbott!supybot@nat/redhat/x-tdjlnbpbuiojizrf> has joined #wildfly-swarm19:15
dmlloydbobmcw: btw, I've tagged & released 1.5.0.CR1 of modules19:29
dmlloydmight go to final within a day or so unless some other hobgoblins appear19:29
bobmcwI assume in the tarpit of repo.jboss.org?19:29
dmlloydyeah it should make its way to central though19:30
dmlloydit's not there yet...19:30
*** lance|afk is now known as lanceball19:35
bbrowninglanceball: I noticed that with booker the username doesn't always display19:35
bbrowningon the account page19:35
lanceballbbrowning: yeah, I noticed that as well19:36
bbrowningit seems to only display the first time I click on account, get redirected to keycloak to login, login, and go back19:36
bbrowningnext time I click account after being logged in it's not there19:36
lanceballyes this is something I noticed a couple of days ago and need to figure out19:36
bbrowningkk :)19:36
lanceballsomehow the react component's state is not getting updated19:36
lanceballbut not sure why19:36
projectodd-ciProject wildfly-swarm build #400: SUCCESS in 33 min: https://projectodd.ci.cloudbees.com/job/wildfly-swarm/400/19:37
projectodd-ci* Bob McWhirter: Remove integs/ since wildfly-swarm-examples now has tests.19:37
projectodd-ci* Toby Crawley: Add sysprop to dump Arquillian swarmjar for debugging19:37
projectodd-ci* Toby Crawley: Allow setting modules and repos when arq testing [SWARM-205]19:37
projectodd-ci* Toby Crawley: Include additional modules when resolving build dependencies.19:37
jbossbotjira [SWARM-205] expose a way to set additional modules and repos when testing via Arquillian [Done (Done) Task, Major, Toby Crawley] https://issues.jboss.org/browse/SWARM-20519:37
projectodd-ci* Toby Crawley: Add test to verify additional repos and modules in arq build [SWARM-205]19:37
projectodd-ci* Toby Crawley: Expose debugging remote arq container via sysprop19:37
projectodd-ciStarting build #89 for job wildfly-swarm-examples (previous build: SUCCESS)19:38
projectodd-ciStarting build #48 for job wildfly-swarm-examples-uberjar (previous build: SUCCESS)19:40
bbrowningI added some OpenShift 3 instructions to the Booker README if anyone's interested - https://github.com/wildfly-swarm/booker/#openshift19:49
projectodd-ciProject wildfly-swarm-examples build #89: FAILURE in 11 min: https://projectodd.ci.cloudbees.com/job/wildfly-swarm-examples/89/19:49
projectodd-ciBob McWhirter: Move ribbon-secured to use the newly-moved keycloak server.19:49
bbrowningthings are getting closer to actually working properly :)19:50
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has joined #wildfly-swarm19:51
projectodd-ciProject wildfly-swarm-examples-uberjar build #48: FAILURE in 12 min: https://projectodd.ci.cloudbees.com/job/wildfly-swarm-examples-uberjar/48/19:52
projectodd-ciBob McWhirter: Move ribbon-secured to use the newly-moved keycloak server.19:52
*** helio-frota <helio-frota!~heliof@8.80.249.186.araujosat.com.br> has joined #wildfly-swarm19:57
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has quit IRC (Ping timeout: 240 seconds)19:57
*** agoncal <agoncal!~agoncal@212-198-109-24.rev.numericable.fr> has quit IRC (Remote host closed the connection)20:01
bobmcwdangit20:01
bobmcwro14nd: thanks for the docker-maven-plugin20:03
* bobmcw assumes you're rhuss20:03
ro14ndthanks ;-) glad that its somewhat useful (started as a spare time project, like jolokia :)20:07
bobmcwro14nd: can I register some other flavor of external image config provider, besides the .props one?20:07
ro14ndyep, shouldn't be hard (at least it was designed that way :)20:08
bobmcwwhat's the method?  a plugin <dependency> and some META-INF/services/?20:08
ro14ndthats the entry point --> https://github.com/rhuss/docker-maven-plugin/blob/master/src/main/java/org/jolokia/docker/maven/config/handler/ExternalConfigHandler.java20:09
ro14ndlemme check ...20:09
ro14ndbobmcw: it's a component in plexus configuration --> https://github.com/rhuss/docker-maven-plugin/blob/mastersrc/main/resources/META-INF/plexus/components.xml20:11
ro14ndwhich has the role "org.jolokia.docker.maven.config.handler.ExternalConfigHandler"20:11
ro14ndrusty, rusty ... (like maven itself :)20:11
bobmcwcan I augment that w/o rebuilding your plugin?20:12
bobmcwI'm pretty ignorant about plexus20:12
ro14ndprobably not ;-( (except if plexus has an extension point here which I don't know)20:12
bobmcw'k, I'll poke around20:12
ro14ndplexus is really only used in Maven anymore and quite, hmm, ancient.20:13
ro14nddo you have a special use case in mind ?20:13
jbossbotgit [wildfly-swarm-examples] push master 543047d.. Bob McWhirter Correct the dependency for keycloak server.20:16
jbossbotgit [wildfly-swarm-examples] push master 3bff79b.. Bob McWhirter Relight the docker example using the docker-maven-plugin....20:16
jbossbotgit [wildfly-swarm-examples] push master URL: http://github.com/wildfly-swarm/wildfly-swarm-examples/compare/9fe4e1a...3bff79b20:16
projectodd-ciStarting build #90 for job wildfly-swarm-examples (previous build: FAILURE -- last SUCCESS #87 6 hr 35 min ago)20:16
projectodd-ciYippee, build fixed!20:30
projectodd-ciProject wildfly-swarm-examples build #90: FIXED in 14 min: https://projectodd.ci.cloudbees.com/job/wildfly-swarm-examples/90/20:30
projectodd-ci* Bob McWhirter: Correct the dependency for keycloak server.20:30
projectodd-ci* Bob McWhirter: Relight the docker example using the docker-maven-plugin. Hidden behind20:30
*** projectodd-ci <projectodd-ci!~PircBotx@ec2-54-145-107-102.compute-1.amazonaws.com> has quit IRC (Remote host closed the connection)20:30
*** jclingan <jclingan!~jclingan@redhat/jclingan> has quit IRC (Quit: My Mac has gone to sleep. ZZZzzz…)20:31
bobmcwrestarting CI20:32
bobmcwinstalled some dockery plugins20:32
bobmcwlet's see if they notice20:32
bobmcwro14nd: eh, I don't honestly know.  But -swarm.jar -> docker could be pretty straightforward20:32
*** projectodd-ci <projectodd-ci!~PircBotx@ec2-54-145-107-102.compute-1.amazonaws.com> has joined #wildfly-swarm20:32
bobmcwand if we can remove boilerplate20:32
bobmcwand possible know which ports to open based on $magic or app's <dependency> list etc20:32
bobmcw<type>swarm</type> and have the full configuration happy would be lovely20:33
bobmcwbut beyond "magic" I have no concrete ideas yet20:33
*** jclingan <jclingan!~jclingan@redhat/jclingan> has joined #wildfly-swarm20:33
*** projectodd-ci <projectodd-ci!~PircBotx@ec2-54-145-107-102.compute-1.amazonaws.com> has quit IRC (Client Quit)20:36
*** projectodd-ci <projectodd-ci!~PircBotx@ec2-54-145-107-102.compute-1.amazonaws.com> has joined #wildfly-swarm20:36
*** rbenevides <rbenevides!~rbenevid@redhat/jboss/RafaelBenevides> has quit IRC (Quit: Quit)20:43
ro14ndyes, that could be quite useful to grab as many meta information as possible for a swarm jar or the swarm build.20:49
bobmcwe.g., wildlfy-swarm-plugin already has (possibly) jboss.http.port, which we could shuffle to a <port> or whatnot20:51
bobmcwplus we know the name of the artifact, no reason to have to specify it, for instance20:51
ro14ndanother crazy idea: Implement an cli option which transforms a -swarm to a docker image (e.g. java -jar myapp-swarm.jar --dockerize). This wouldn't add much (in percent its less than 1% ;) since only needs to create a simple tar with the -swarm jar + Dockerfile and send it via HTTP to the Docker daemon.20:52
bobmcwdescriptorRef of "artifact" doesn't pick up the -swarm classifier20:52
bobmcwro14nd: Oh, that'd be pretty neat-o20:52
bobmcwJDK can do tar already?20:52
ro14nd@bobmcw: what we found incredible difficult  is to share maven plugin configuration between different plugins. maybe its better to extract meta data from the -swarm artefact.20:53
ro14ndbobmcw: good point. tbh I don't know.20:53
bobmcwyah, across plugins is sorta goofy20:53
bobmcwI've managed when I own both plugins20:53
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has joined #wildfly-swarm20:53
bobmcwbut it's quite a hack, I think20:53
bobmcwhttps://github.com/wildfly-swarm/wildfly-swarm/blob/master/plugin/src/main/java/org/wildfly/swarm/plugin/maven/MultiStartMojo.java#L260-L28320:54
bobmcwthat's where my :multistart goes and mucks around in a :start <execution>20:54
bobmcwread-only20:54
bobmcwthough, we could subclass/extend your docker plugin, and have a DockerMojo in -swarm that does magic.20:55
bobmcwbut I like your --dockerize20:55
ro14ndbobmcw: https://code.google.com/p/jtar/downloads/detail?name=jtar-1.1.jar is 12 kb though. tar should be no showstopper.20:55
jbossbotTitle: jtar-1.1.jar - jtar - Release 1.1 - Java Tar library - Google Project Hosting20:55
bobmcwro14nd: excellent, I may pursue that20:55
bobmcwyour :start and :stop are also nice20:56
* bobmcw has much respect for people who write (and document!) Mojos20:56
*** os890 <os890!~chatzilla@213-33-15-102.adsl.highway.telekom.at> has joined #wildfly-swarm20:59
ro14ndbobmcw: wrt to descriptorRef: You are right, have to look at it. One probably have to write an own assembly descriptor (which btw can be done also inline).21:00
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has quit IRC (Ping timeout: 265 seconds)21:00
bobmcwyah, that's what I ended up with21:00
bobmcwhttps://github.com/wildfly-swarm/wildfly-swarm-examples/blob/master/docker/docker-jaxrs/pom.xml#L76-L13321:00
jbossbotgit [wildfly-swarm-examples] push master 14a98c2.. Bob McWhirter Adjust some READMEs.21:02
jbossbotgit [wildfly-swarm-examples] push master URL: http://github.com/wildfly-swarm/wildfly-swarm-examples/commit/14a98c2e821:02
ro14ndnice trick mangle the maven config. it's a pity that this isn't support directly (sharing of plugin configs,)21:02
projectodd-ciStarting build #91 for job wildfly-swarm-examples (previous build: FIXED)21:03
bobmcwfwiw, I stole that from https://github.com/TimMoore/mojo-executor21:03
ro14ndbobmcw: btw why do you chose "entryPoint" over "cmd" ?21:04
projectodd-ciStarting build #49 for job wildfly-swarm-examples-uberjar (previous build: FAILURE -- last SUCCESS #46 7 hr 26 min ago)21:05
bobmcwbecause I'm docker-ignorant, and was cribbing from https://github.com/wildfly-swarm/wildfly-swarm-examples/blob/543047d871d612e182bbfc1a146fb172bfe5331e/docker-jaxrs-app/Dockerfile#L621:05
bobmcw(which I think kenfinnigan wrote initially)21:05
ro14ndhave to check why a dependencySet in the assembly doesn't work for artifacts with classifiers. would look cleaner than.21:05
bobmcwbtw, if I tried putting that Dockerfile under src/main/docker/ and used dockerFileDir, I couldn't ../../../target/thing-swarm.jar21:05
bobmcwerror being outside of the context21:06
bobmcwI think it would've just grabbed example.jar, and not the example-swarm.jar built and attached by the other plugin21:06
bobmcwartifactRef=artifact didn't know to look for -swarm, since it's not codified anywhere21:06
bobmcwerr, descriptorRef21:06
bobmcwwe could create a ref that was for swarmArtifact mayhap21:07
bobmcwanyhow, it works as-is, so now we can figure out cleaner :)21:07
ro14ndbobmcw: If you want to have usage semantic like 'docker run my-swarm-image <args to the swarm-jar>' then entry point is the right choice (e.g. when you want to allow extra arguments when starting the docker image). However this is normally used for CLI kind of apps only (services typicyally evaluate env vars).21:07
kenfinniganbobmcw: i did, but it was also kind of cribbed from someone else ;-)21:07
bobmcwyah, we do want to allow args21:08
bobmcw-= THIS MESSAGE NOT LOGGED =-21:08
bobmcwor I learn more about docker and figure out ENV_VARS usage21:08
ro14ndentrypoint makes it harder to bash into the image. i.e. "docker run myswarm-image bash " doesn't work then (you would need to do a docker --entrypoint bash myswarm-image21:08
bobmcwah21:08
ro14ndif you want to run in Kubernetes / OpenShift env vars are the canonical way.21:09
ro14nd(because you can specify them in a DeploymentConfig ....)21:09
bobmcw'k, I'll lean on bbrowning21:09
bobmcwro14nd: thanks for your help; I'll stab around some more and see what I can work up.21:11
tcrawleyhey, speaking of -Djboss.http.port, should: mvn package -Djboss.http.port=8100 give me a swarmjar that boots undertow on 8100?21:11
ro14ndbobmcw: yes, ADD in a Dockerfile works only for subdirs (and .). thats a docker limitation. (sorry, I'm slow in reading and writing. its late ;-)21:11
tcrawleybecause it doesn't seem to, at least in the jaxrs-war example21:11
bobmcwwhen I made ${project.basedir} my dockerFileDir, it seemed to add everything to the resulting .tar21:11
bobmcwnot just the one ADD line bit21:11
ro14ndbobmcw: ok. thats interesting.21:12
bobmcwif there was a <dockerFile> config, which didn't add everything but just built from the Dockerfile, that'd be good21:12
bobmcwdocs do say "Any additional files located in this directory will also be added to the image."21:13
bobmcwso I wasn't surprised21:13
bobmcwbut some flavor/flag/other-config that didn't do that, and we could go back to a regular Dockerfile21:13
bobmcwwhich I think might be ideal21:13
bobmcwanyhow, it's beer o'clock here now, I'll pick it up in the morning21:14
bobmcwro14nd: thanks!21:14
ro14ndI will have a look. the direct Dockerfile support will be enhanced soon btw.21:14
ro14ndsee you ...21:14
projectodd-ciYippee, build fixed!21:17
projectodd-ciProject wildfly-swarm-examples-uberjar build #49: FIXED in 12 min: https://projectodd.ci.cloudbees.com/job/wildfly-swarm-examples-uberjar/49/21:17
projectodd-ci* Bob McWhirter: Correct the dependency for keycloak server.21:17
projectodd-ci* Bob McWhirter: Relight the docker example using the docker-maven-plugin. Hidden behind21:17
projectodd-ci* Bob McWhirter: Adjust some READMEs.21:17
projectodd-ciProject wildfly-swarm-examples build #91: SUCCESS in 14 min: https://projectodd.ci.cloudbees.com/job/wildfly-swarm-examples/91/21:17
projectodd-ciBob McWhirter: Adjust some READMEs.21:17
jbossbotgit [wildfly-swarm-examples] push master fff885d.. helio-frota some readme adjusts21:36
jbossbotgit [wildfly-swarm-examples] push master deb3284.. Bob McWhirter Merge pull request #43 from helio-frota/README-ADJUST...21:36
jbossbotgit [wildfly-swarm-examples] push master URL: http://github.com/wildfly-swarm/wildfly-swarm-examples/compare/14a98c2...deb328421:36
projectodd-ciStarting build #50 for job wildfly-swarm-examples-uberjar (previous build: FIXED)21:37
projectodd-ciStarting build #92 for job wildfly-swarm-examples (previous build: SUCCESS)21:39
*** jclingan <jclingan!~jclingan@redhat/jclingan> has quit IRC (Read error: Connection reset by peer)21:43
*** kenfinnigan <kenfinnigan!~ken@redhat/jboss/kenfinnigan> has quit IRC (Quit: Leaving.)21:49
projectodd-ciProject wildfly-swarm-examples-uberjar build #50: SUCCESS in 14 min: https://projectodd.ci.cloudbees.com/job/wildfly-swarm-examples-uberjar/50/21:51
projectodd-ci00hf11: some readme adjusts21:51
projectodd-ciProject wildfly-swarm-examples build #92: SUCCESS in 14 min: https://projectodd.ci.cloudbees.com/job/wildfly-swarm-examples/92/21:53
projectodd-ci00hf11: some readme adjusts21:53
bobmcwhelio-frota: thanks for the updates21:55
helio-frotabobmcw, np :]21:56
helio-frotabobmcw, adjusting now the clojure example21:56
helio-frotato do a PR21:56
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has joined #wildfly-swarm21:56
bobmcwalso note that we're adjusting the note at the headers as to where to submit bugs etc21:57
bobmcwthe datasources ones have the latest header ref'ing JIRA21:57
helio-frotayup21:57
helio-frotaand readme.md to readme.adoc plus syntax ofc21:58
bobmcwthere ya go21:58
bobmcwmuch appreciated21:58
bobmcwfigure out gitbooks are .adoc, so let's stick to that for everything we can21:58
bobmcwro14nd: I feel like your <port> mapping stuff would be useful for boot2docker on my mac, but I don't quite follow what I'd do to get my docker-machine's IP address jammed into a variable22:10
bobmcw(based on the shootout readme)22:11
bobmcw$(docker-machine env default) shell will give me the IP of it22:11
bobmcwbut then I'm stumped on the integration/config that'll work with native docker and boot2docker22:12
*** tcrawley is now known as tcrawley-away22:13
bobmcwor docker-machine ip default22:14
helio-frotanot in one shot...22:15
helio-frotaclojure example need some tweak22:15
helio-frotato work with IT tests22:16
helio-frotai use this to run: mvn clean clojure:compile package wildfly-swarm:run22:16
bobmcwlike the scala example, I'm sure you can get clojure:compile to run during the compile phase of maven22:16
helio-frotatesting...22:19
*** mgoldmann is now known as mgoldmann|away22:19
*** bbrowning is now known as bbrowning_away22:21
jamezpI don't see one, but is there a JIRA to add batch (JSR 352) to WF Swarm?22:24
bobmcwjamezp: probably not, needs to be added to config-api first, but file an @issue22:25
proddbotWe'd be tickled pink if you would file an issue at https://issues.jboss.org/browse/SWARM22:25
bobmcw(and submit a PR!)22:25
bobmcwI'm out for the night, for serious this time.22:26
helio-frotait seems that is compiling but still getting 404 in tests...22:26
*** os890 <os890!~chatzilla@213-33-15-102.adsl.highway.telekom.at> has quit IRC (Ping timeout: 240 seconds)22:26
jamezpAh. I thought everything was already in the config-apu. I'll have a quick look. I started adding some things already :)22:26
helio-frotagoing to investigate more..22:26
bobmcwjamezp: there's a .json in config-api to control what subsystems we're doing22:26
bobmcwand we've probably not added them all yet22:26
bobmcwfeel free to add'em all22:27
jamezpAh okay... ...now really get out of here bobmcw :)22:27
bobmcwso far they've been added when we're pulling them from -swarm, which is not the complete set yet22:27
*** bobmcw <bobmcw!~bobmcw@redhat/jboss/bobmcw> has quit IRC (Remote host closed the connection)22:28
ytanabetcrawley-away, Thank you for SWARM-205 ! Now I can do arq test with postgresql.22:29
jbossbotjira [SWARM-205] expose a way to set additional modules and repos when testing via Arquillian [Done (Done) Task, Major, Toby Crawley] https://issues.jboss.org/browse/SWARM-20522:29
*** bgeorges_ <bgeorges_!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has joined #wildfly-swarm22:36
*** bgeorges <bgeorges!~bgeorges@115-188-52-145.jetstream.xtra.co.nz> has quit IRC (Read error: Connection reset by peer)22:37
helio-frotajust remembered this example worked only with 'war' not 'jar' :D22:40
helio-frotaweirdo but generates a runnable jar :]22:42
helio-frotagoing to send a PR22:42
*** lanceball is now known as lance|afk22:49
*** bobmcw <bobmcw!~bobmcw@va-67-233-86-120.dhcp.embarqhsd.net> has joined #wildfly-swarm23:28
*** bobmcw <bobmcw!~bobmcw@va-67-233-86-120.dhcp.embarqhsd.net> has quit IRC (Changing host)23:28
*** bobmcw <bobmcw!~bobmcw@redhat/jboss/bobmcw> has joined #wildfly-swarm23:28
*** ChanServ sets mode: +o bobmcw23:28
*** bobmcw <bobmcw!~bobmcw@redhat/jboss/bobmcw> has quit IRC (Ping timeout: 246 seconds)23:35
*** helio-frota <helio-frota!~heliof@8.80.249.186.araujosat.com.br> has quit IRC (Quit: Leaving)23:40

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!