Log of the #fcrepo channel on chat.freenode.net

Using timezone: Eastern Standard Time
* github-ff joins00:50
[fcrepo-module-auth-webac] acoburn opened pull request #39: add filesystem/classpath backstop (master...fcrepo-1792) http://git.io/vW9bN
* github-ff leaves
* dwilcox leaves03:38
* awead joins06:28
* awead_away leaves06:29
* awead leaves07:33
* awead joins07:35
* awead leaves07:48
* awead joins07:52
* acoburn joins08:36
* dhlamb joins08:48
* awead leaves08:59
* jrgriffiniii leaves09:24
* awead joins09:27
* whikloj joins09:34
[WebAC] Standup - early edition09:35
Worked on fcrepo-1795 - MessageBodyWriter does not seem to stack, looking at MessageInterceptor09:36
Worked on fcrepo-1794 - PR for ontology, but some questions regarding acoburn's implementation in the ticket
Today look further into fcrepo-1760, try to find where the 404 is originating.09:37
* whikloj leaves
* umgrosscol joins09:51
* mohamedar joins09:52
* ajs6f joins09:56
* jgpawletko joins09:57
* bseeger joins09:58
* peichman joins10:09
* whikloj joins10:17
* whikloj leaves10:18
* awoods leaves10:20
* awoods joins10:21
* awead leaves10:28
* dhlamb leaves10:46
* awead joins10:54
<awoods>peichman: https://plus.google.com/hangouts/_/event/c1glu6soq43r1rr6ou17qtobug811:01
https://wiki.duraspace.org/display/FF/Semantics+of+ACL+Modes11:05
* chadmills joins
<mohamedar>SPRINT STANDUP11:10
Working on FCREPO-1797, Created scripts to backup and test restoring fcrepo.home11:12
* awead_ joins11:31
* awead leaves
* bseeger leaves11:35
* dhlamb joins11:41
* dwilcox joins11:42
* umgrosscol leaves11:54
* escowles joins12:04
* dwilcox leaves12:16
* bseeger joins12:20
* travis-ci joins12:31
jrgriffiniii/fcrepo4-client#1 (fcrepo-1443-jrgriffiniii - 31aa509 : griffinj@lafayette.edu): The build passed.
Change view : https://github.com/jrgriffiniii/fcrepo4-client/compare/336f6c0c9cf7...31aa509dbd0f
Build details : https://travis-ci.org/jrgriffiniii/fcrepo4-client/builds/87928646
* travis-ci leaves
* whikloj joins12:34
* chadmills leaves
* dwilcox joins
* ajs6f leaves12:37
<awoods>peichman: I have things working with the fcrepo-webapp-plus picking up local fcrepo-module-auth-webac changes... let me know when you are around.
<whikloj>awoods/acoburn: Sorry I missed the call, not sure i understand how hangouts works on my phone :(12:39
<awoods>whikloj: we saw that you were trying to call on a different hangout
<whikloj>awoods: gah, darn I have about 4 different "ongoing calls" in the app
<awoods>whikloj: anyways, welcome back.12:40
whikloj: the main actions/take-aways from the call were:
whikloj: Hold off on this ticket for now: https://jira.duraspace.org/browse/FCREPO-1794
whikloj: we will discuss Modes again on Friday and decide if we will include the above ticket in this sprint.12:41
<whikloj>awoods: ok, is there a particular reason to hold off?
<awoods>whikloj: it is unclear at this point if the use cases require more Modes that have already been implemented.12:42
<whikloj>awoods: ok
awoods: I'll move to fcrepo-1760 and check the VersionRdfContext you referenced
<awoods>whikloj: Great. ..and, if we a trying to wrap up implementation this week, it is important to focus on the tickets that are more on the critical path.12:43
<whikloj>awoods: do you have something in that vein?12:44
* dwilcox leaves12:45
<awoods>whikloj: 1760 is a definite need
<whikloj>awoods: ok12:46
* osmandin joins12:56
* dwilcox joins12:59
* peichman1 joins
* peichman leaves13:02
<awoods>escowles: https://plus.google.com/hangouts/_/event/c1glu6soq43r1rr6ou17qtobug8
<escowles>awoods: i'll be right there
* mikeAtUVa joins13:05
<awoods>welcome aboard mikeAtUVa!
* ajs6f joins13:08
* dwilcox leaves
* avmich joins13:13
* cbeer joins13:19
ajs6f: thanks for your reply. i think the possible objections you suggested are similar to the ones I was going to raise. i agree there's a tension between an object repository and an RDF repository, and maybe I'm more sympathetic to the RDF side.13:22
<ajs6f>cbeer: That's certainly a reasonable POV. My only objection would be that I think it misleading to call that kind of RDF-centric product "Fedora".
cbeer: We could call it, "Son of Fedora".13:23
<escowles>rdfedora
<ajs6f>There must be some way to visually combine the Fedora "meatballs" and the RDF logo.13:24
<cbeer>I guess my other objection is: what do other LDP implementations do, and can we ensure some level of interoperability?... of course, i think the answer to that is "it depends" and "no"
<ajs6f>cbeer: I dunno. I agree about the first question, but I think the answer to the second is, "If Fedora is restrictive, then people can go from Fedora elsewhere, but not neccessarily the other way." and that sound fine to me.13:25
<cbeer>if we make that message clear, that works for me. hopefully tjohnson's excellent rdf-ldp work will make that easy / easy enough.13:26
<ajs6f>cbeer: Cool. I had (months ago) a really good email conv with tjohnson about this triplestore vs. object repo tension. I think we can make that sort of thing plain if we work a little at it.13:27
<cbeer>+1
<ajs6f>Sounds like a job for awoods and dwilcox.13:28
<awoods>on a call
<ajs6f>Strange, that's not like awoods.
* peichman1 leaves13:30
* peichman joins13:36
@awoods: I figured out my issue, it wasn't a build problem, it was that the FAD is never called when the authenticated user is a fedoraAdmin13:37
<awoods>good
<peichman>so now I have a better idea of where I need to hook in to do the delegated user permission check
<awoods>on a call13:38
<avmich>hi, I've got the following diagnostics when running Fedora 4.2.0 -13:39
ERROR 17:38:31.772 (SystemNamespaceRegistry) Node types were read from the system content, and appear to be inconsistent or invalid: repo
what does it mean?13:40
it seems like Fedora doesn't correctly start, and this is the first indication of problems in console logs
<ajs6f>Have you tinkered with the node definition file?13:41
<avmich>I think no. I don't know what is node definition file.13:44
I've modified Infinispan, ModeShape and JGroups config files
but node definition file...
<ajs6f>avmich: What did you do to the ModeShape config file?
<avmich>changed in fcrepo-configs/src/main/resources/config/clustered/repository.json13:46
"cacheConfiguration" : "${fcrepo.ispn.configuration:config/infinispan/leveldb/infinispan.xml}",
to
"cacheConfiguration" : "${fcrepo.ispn.configuration:config/infinispan/clustered/infinispan.xml}",
I've also changed fcrepo-webapp/src/main/resources/spring/repo.xml
<ajs6f>avmich: What change did you make there?13:47
<avmich>from p:repositoryConfiguration="${fcrepo.modeshape.configuration:classpath:/config/minimal-default/repository.json}"/>
to p:repositoryConfiguration="${fcrepo.modeshape.configuration:classpath:/config/clustered/repository.json}"/>
so just changing pointers to files
I also changed that fcrepo-configs/src/main/resources/config/infinispan/clustered/infinispan.xml file :) but this is already in Infinispan area, right?13:48
<ajs6f>avmich: Yes, seems reasonable. Can you turn up the logging for org.modeshape? It probably won't show anything, but it's worth trying.
<avmich>how?
<ajs6f>There should be a logging.xml file somewhere near your other config.13:49
<avmich>let me see...
no such file in the whole tree...13:50
<ajs6f>try logback.xml
<avmich>aha, I see a few - one is in fcrepo4/fcrepo-webapp/src/main/resources/logback.xml . I guess that's get copied to several places in fcrepo4/fcrepo-webapp/target ?..13:52
aha, all identical13:53
fcrepo4/fcrepo-webapp/target/classes/logback.xml , fcrepo4/fcrepo-webapp/target/fcrepo-webapp-4.2.0/WEB-INF/classes/logback.xml , fcrepo4/fcrepo-webapp/target/console-work/WEB-INF/classes/logback.xml13:54
where to change?
<ajs6f>There should be only one. Are you looking in the source tree? No, look in the actual artifact you are deploying.13:55
<avmich>hm13:57
I'm using Maven, running it in fcrepo4/fcrepo-webapp directory
mvn jetty:run13:58
<ajs6f>Okay, so fcrepo-webapp/target is your artifact.
<avmich>which file it uses?
<ajs6f>No, the whole directory.
<avmich>of, it has 3 logback.xml files
<ajs6f>Where are they
<avmich>fcrepo4/fcrepo-webapp/target/classes/logback.xml , fcrepo4/fcrepo-webapp/target/fcrepo-webapp-4.2.0/WEB-INF/classes/logback.xml , fcrepo4/fcrepo-webapp/target/console-work/WEB-INF/classes/logback.xml13:59
probably not the last one?..
will come back in ~20 minutes :(14:00
<ajs6f>Hard to say. You'll need to check the Jetty-maven docs for that. Otherwise, I would try changing the one in the fcrepo-webapp source tree and rebuilding.14:01
<avmich>thank you. Let me try.14:31
wonder which one is logging for org.modeshape... I have logger names org.fcrepo.auth , org.fcrepo.connector.file , org.fcrepo.http.api , org.fcrepo.http.commons , org.fcrepo.jms , org.fcrepo.kernel , org.fcrepo.transform , org.fcrepo14:39
<whikloj>avmich: I'd try kernel14:40
<ajs6f>avmich: No, it's just org.modeshape.
Those are all just package names. Java logging normally is managed by package.
<avmich>should I add a logger?..
<ajs6f>avmich: Yes.
avmich: This may or may not work, but it's worth a try.14:42
<avmich>adding <logger name="org.modeshape" additivity="false" level="${modeshape.log:-INFO}"> <appender-ref ref="STDOUT"/> </logger> before <root...14:43
<ajs6f>You can just replace ${modeshape.log:-INFO} with DEBUG.
<avmich>is it a good enough level?
ah, ok
<ajs6f>avmich: Also, once you build and run with this, you can find / grep the various logback.xml files in target to see which one ended up with this modification. Then you don't have to keep fooling with the source code. You can just alter the artifact in deployment.14:45
<avmich>if they are not all copied from the same place...14:46
<ajs6f>avmich: If they are, you're no worse off than you are now. This is one reason that running jetty-maven is a pretty limited way to tinker.
<avmich>which way is better?
<ajs6f>avmich: What do you mean? Which way of what?14:47
avmich: Better in what way?
<avmich>instead of jetty-maven... How it is to launch?
which way to tinker is better?14:48
* dwilcox joins
<ajs6f>avmich: That depends on what you are doing. E.g. if you aren't tinkering with the Java code itself, just unpack a WAR in a servlet container like Tomcat or Jetty and you can tinker with the files without having to worry abot Maven, which is fundamentally a build tool.14:49
<avmich>ah, you mean Maven isn't convenient because it rebuilds everything every time you want to launch? And when you work with Jetty directly, you can just update a config file and re-launch, without rebuilding?14:50
<ajs6f>avmich: Unless you are actually building code (and it doesn't sound at all like you are) Maven isn't go to help you in any way.
Maven is a build tool.
<avmich>I'm rebuilding Fedora 4.2.0 from a git tag state14:52
<ajs6f>avmich: Why? Why not just download a 4.2 webapp?
<avmich>because I'm changing config files which go inside those WARs :)14:54
<ajs6f>avmich: That doesn't matter at all. You can just unpack the WARs and change whatever you want.
WARs are just special forms of TAR files.
Use the Java `jar` command.14:55
<avmich>right... I also wanted to look inside. So I rebuilt Fedora for other reasons. I can re-assemble WARs, just didn't get to it - have to write a script which would unpack, change config and pack again14:57
<ajs6f>avmich: No you do not.
avmich: Any common servlet container can run a webapp from an unpacked directory.
<acoburn>avmich: when you say you're changing configuration files, which files do you mean?14:58
avmich: repository.json?
<ajs6f>avmich: And most can monitor a directory for changes and restart the webapp housed therein, so you don't even have to restart the server.
<avmich>ok, so I can get WARs from Fedora distribution, unpack them, update configs and just launch in unpacked state?
<acoburn>avmich: or the spring files?
<ajs6f>avmich: Yes, you can and should. Unless you are working on code, you have use for Maven.
<avmich>infinispan.xml , repo.xml, repository.json and that jgroups-tcp.xml , don't remember exact name...
<ajs6f>avmich: There is no distinction to the servlet container between those files. They're all just files in a directory.14:59
<acoburn>avmich: because if I'm not mistaken, everything you're trying to do can be done by setting environmental variables
* dwilcox leaves
* cbeer leaves
<ajs6f>avmich:acoburn: That may very well be. Even easier than unpacking a WAR.
<acoburn>avmich: there's no need to touch the code to change those
<avmich>I'm setting several when I issue mvn jetty:run
<ajs6f>avmich:acoburn: I think it's pretty clear now that you should not be using mvn jetty:run. It's just making your life difficult.15:00
<acoburn>personally, I avoid it at all costs
<avmich>btw, regarding repo.xml file - what this line mean?15:01
p:repositoryConfiguration="${fcrepo.modeshape.configuration:classpath:/config/clustered/repository.json}"/>
specifically, why there are two colons?15:02
<acoburn>avmich: here's a snippet from my /etc/sysconfig/tomcat file: https://gist.github.com/acoburn/a61d1dde2e47ba5f6937
<avmich>personally, I'm always trying to avoid Maven. But instructions are written this way.
<acoburn>avmich: it allows me to control the config files in a directory outside of the deployed fedora system
avmich: which you should be doing in production if you plan to modify any of them15:03
<ajs6f>avmich: What instructions?>
<avmich>I was trying to understand the way those $-prefixed strings work - and that's why I decided, being on the safe side, to just edit config files instead of setting parameters at launch time
<acoburn>avmich: that's all spring syntax
avmich: it means "take the value from such-and-such system property"15:04
<ajs6f>avmich: They select from a supplied system property or in its absence, a default.
<acoburn>avmich: but if it's not available use what follows the colon
<avmich>didn't check Spring docs... but this format is used elsewhere. Is it always read by Spring code?
<ajs6f>avmich: When its in a Spring XML file it is.
<avmich>ok, but why there are two colons?
<ajs6f>avmich: Why wouldn't there be? What is inherently wrong about that?15:05
<acoburn>avmich: the first colon delimits the system property from the default
<ajs6f>avmich: One separates the proerpty name and default value. The other is a part of the default value.
<acoburn>avmich: the default happens to contain a colon
<ajs6f>acoburn: I'm leaving this to you. It's getting confusing listening to each or us talk at the same time.
<acoburn>avmich: in spring syntax, classpath: is a url scheme
<avmich>aha... thanks15:06
<acoburn>avmich: another value might use file: or http:
<avmich>found where I got Maven instructions from: https://github.com/fcrepo4/fcrepo4/tree/fcrepo-4.2.015:07
so should I switch to Tomcat now? :)15:08
<acoburn>avmich: yes, that's certainly a "quick and dirty" way to get fcrepo up and running
<ajs6f>avmich: Those are instructions specifically for building and running from source.
<avmich>I'd prefer Jetty just to save the switching...
<acoburn>ajs6f++
<ajs6f>https://wiki.duraspace.org/display/FEDORA40/Deploying+Fedora+4+Complete+Guide15:09
acoburn: There should probably be a link at the top of the source tree README saying "If you just want to run Fedora, go here and read this."
<acoburn>ajs6f++15:10
<ajs6f>acoburn: avmich was tricked.
<acoburn>ajs6f: I see that the deployment guide ^^^ mentions jetty:run as good _for testing_
<ajs6f>acoburn: That should be clarified. Testing _code_, not config.
<avmich>our story is like this. We got some version of Sufia and developed a service on top of that. Then we wanted to have a cluster. I'd prefer to use the latest versions of software, but seems like Fedora 4.4 and 4.3 have clustering broken (so far?) . And our Sufia had Fedora 4.1 . Now I'm trying to get 4.2 working. So we're working with different Fedoras...
<acoburn>ajs6f: I'll add a link to this in the readme15:11
<ajs6f>acoburn++
<bseeger>acoburn: here, too, https://wiki.duraspace.org/display/FF/Guide+for+New+Developers#GuideforNewDevelopers-Tomcat
<acoburn>bseeger++15:12
<ajs6f>bseeger: You can change that yourself, or if not, we will fix it so that you can.
<avmich>saw all that... would you really recommend Tomcat?
<ajs6f>avmich: Sure, or Jetty, or whatever servlet container you like.
<acoburn>avmich: I use tomcat, but others run fcrepo in jetty just fine15:13
<avmich>that's the question. So, no - I can stay with Jetty. Ok.
<acoburn>avmich: just not mvn jetty:run
<avmich>yes, I see
<ajs6f>avmich: Run your fingernails through your mind and rip out any trace of mvn jetty:run.
<avmich>I worked both with Tomcat and Jetty, Jetty more recently, that's why I slightly prefer that... seems like a leaner system15:14
with great preasure
pleasure
<bseeger>ajs6f: sure, I'll can change it later. Is it as simple as putting in a note about the preference for tomcat?15:15
<ajs6f>bseeger: If you are logged in, you should see an "Edit" button in the top right?
Or link, or something.
<bseeger>ajs6f: yes15:16
<ajs6f>bseeger: Cool.
* awead joins15:23
* awead_away leaves15:26
<avmich>ok, all logback.xml are identical after build15:34
* github-ff joins15:59
[fcrepo4] acoburn opened pull request #927: clean up README files (master...update-readme) http://git.io/vWFmv
* github-ff leaves
<acoburn>avmich: that's for you ^^^16:00
avmich: here's what it looks like: https://github.com/acoburn/fcrepo4/tree/update-readme
* bseeger leaves16:10
* travis-ci joins16:16
fcrepo4/fcrepo4#4105 (master - 8cceaeb : Bethany Seeger): The build was broken.
Change view : https://github.com/fcrepo4/fcrepo4/compare/1e4d09d30d70...8cceaebd0121
Build details : https://travis-ci.org/fcrepo4/fcrepo4/builds/87759854
* travis-ci leaves
* ajs6f leaves16:18
* dhlamb leaves16:20
* avmich leaves16:23
* cbeer joins16:28
* osmandin leaves16:29
* dwilcox joins
* ajs6f joins16:33
* avmich joins16:41
acoburn: thank you
<acoburn>awoods: ping16:43
* ajs6f leaves16:44
<acoburn>avmich: I'm glad you think that's better, thanks for motivating us to get our act together on that README file :-)16:45
<avmich>it's probably never perfect though :)
I might field-test it...16:46
* jgpawletko leaves
<acoburn>avmich: that would be great. documentation is often the thing that gets left behind
<avmich>yes :(16:47
ok, now I'm launching like this - java -D... -D... -jar <fedora.jar> --jetty_options...16:57
* mikeAtUVa leaves
<avmich>the error is ERROR 19:36:13.873 (SystemNamespaceRegistry) Node types were read from the system content, and appear to be inconsistent or invalid: repo16:58
as for extra logging...
hmmmm, don't see anything suspicious... right before the error I have16:59
DEBUG 19:36:13.771 (RepositoryCache) Found existing 'system' workspace in repository 'repo'
<acoburn>avmich: this is the "one click" application?
<avmich>I'm launching two instances of Fedora on the same host - in two directories and two JVMs, with different Jetty ports... and corresponding cluster configs...
yes, this is -jar ./fcrepo-webapp/target/fcrepo-webapp-4.2.0-jetty-console.jar17:00
<acoburn>and are you setting -Dfcrepo.home
<avmich>let me check...17:01
no, not this one
<acoburn>that's the key thing to add
<avmich>is it mentioned somewhere?
<acoburn>if by "somewhere" you mean the wiki, yes, but that may not be immediately obvious17:02
the error you're getting is suggesting that the second instance is reading the data directory from the first instance
<avmich>yes...
I see - https://wiki.duraspace.org/display/FEDORA4x/Deploying+Fedora+4+Complete+Guide#DeployingFedora4CompleteGuide-DeployingwithJetty9
it says fcrepo.home tells where the data sits
<acoburn>by default the data directory is set to the directory in which you executed the start command17:03
so, if you've got 2 containers that are starting up in the same directory, just do this:
* dwilcox leaves
<acoburn>java -jar …. -Dfcrepo.home=/var/fedora1 fedora.jar17:04
and
java -jar … -Dfcrepo.home=/var/fedora2 fedora.jar
(or whatever the commands happen to be — the key point is to make the data directories different)
<avmich>aha... I'm launching jetty-console.jar from fcrepo4 directory, that's my current working directory. The fcrepo4-data is within fcrepo-webapp...
no, no, I have two completely separate directories17:05
two copies of Fedora code
hm
<acoburn>and you don't have anything set in JAVA_OPTS?
echo $JAVA_OPTS
<avmich>nothing in JAVA_OPTS17:06
let me cd to fcrepo-webapp before launching Jetty...
aha, different error...17:07
* dwilcox joins
<avmich>oh, used wrong ports...17:08
<acoburn>Are you setting the activemq port for the two fedoras? I'm thinking there may be a conflict
-Dfcrepo.dynamic.jms.port=6161617:09
<avmich>I do
<acoburn>and -Dfcrepo.dynamic.stomp.port=6161317:10
<avmich>I used something like MAVEN_OPTS="-Xmx512m -Djgroups.tcp.address=localhost -Djgroups.tcp.port=7802 -Djgroups.tcpping.initial_hosts=localhost[7801] -Dfcrepo.dynamic.jms.port=61618 -Dfcrepo.dynamic.stomp.port=61615" mvn -Djetty.port=8082 jetty:run
now I'm copying these parameters to Jetty one-click launch command
and I change all these parameters on another instance
so... I don't see that error...17:11
hmmm, I only added modeshape logging and started as Jetty one-click launch, instead of Maven... it seems to go away
<acoburn>another reason not to use mvn jetty:run :-)17:12
<avmich>I also have ERROR 21:09:19.028 (ClusteredChangeBus) Loopback changeset '0c07a45b-973c-41c1-b585-8c13d6637de0' was never received back on 'ClusteringService[cluster_name='modeshape-cluster', address=ip-172-30-0-14-18283]'. Make sure your JGroups configuration uses 'loopback=true' and if applicable 'loopback_separate_thread=true' , but I saw that before...
<acoburn>ok, that gets into unfamiliar territory for me — I haven't done anything with clustering
<avmich>this one doesn't always prevent me to run clustering, so I'm largely ignoring that for now... but where did that one go? About repo?..17:13
is it because I launched from different directory?..
* dwilcox leaves
<avmich>no, Maven launched correctly another instance with similar layout...17:14
hmm
<acoburn>maybe awoods has some thoughts on this? I need to head out17:15
* acoburn leaves17:16
<avmich>ok, thank you! I still have some suspicions
need to check...17:17
* chadmills joins17:23
* dwilcox joins17:32
* peichman leaves17:39
* mohamedar leaves17:42
* whikloj leaves17:53
* FlowRidda joins17:59
* FlowRidda leaves18:01
* github-ff joins18:11
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/vWFxa
fcrepo4/master be7beb8 Aaron Coburn: Clean up README files...
* github-ff leaves
* github-ff joins18:12
[fcrepo4] awoods closed pull request #927: clean up README files (master...update-readme) http://git.io/vWFmv
* github-ff leaves
* chadmills leaves18:15
* travis-ci joins18:27
fcrepo4/fcrepo4#4108 (master - be7beb8 : Aaron Coburn): The build was fixed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/8cceaebd0121...be7beb8b3d2e
Build details : https://travis-ci.org/fcrepo4/fcrepo4/builds/87999445
* travis-ci leaves
<avmich>ok...18:36
if I created a cluster and added a node to it - should I expect the new node to pick up the data from the cluster? Replication mode is configured for Infinispan...18:37
<awoods>avmich: is your "node" a new Fedora server or a resource?18:40
<avmich>a new Fedora instance
same host, but different directory, JVM, ports...
different fcrepo4-data directory
<awoods>avmich: it depends on your cluster configuration, but in replication mode, the new Fedora instance should become populated with the content found on the other clustered instances.18:41
avmich: fyi, the following ticket has been resolved: https://jira.duraspace.org/browse/FCREPO-1739
<avmich>that looks very good... may be I just have to switch to 4.4...18:42
wonder if that would help
<awoods>avmich: help what exactly?
* dwilcox leaves
<avmich>to make it copy data from node to node
replicate
<awoods>avmich: is replication not working with 4.2?18:43
<avmich>yes, at least in my config
<awoods>hmm
<avmich>there are some errors... may be related
like...
ERROR 22:28:40.395 (velocity) Exception rendering #parse(views/common-metadata.vsl) at /views/node.vsl[line 37, column 13]
org.eclipse.jetty.io.EofException: null18:44
at org.eclipse.jetty.io.ChannelEndPoint.flush(ChannelEndPoint.java:192) ~[na:na]
ending with
Caused by: java.io.IOException: Broken pipe
<awoods>avmich: does the HTML view show up in the browser for each of your F4 instances?
<avmich>yes
one has 1 child, another 0
<awoods>avmich: that should not be effecting replication.
<avmich>another one:18:45
ERROR 22:28:40.401 (ServerRuntime$Responder) An I/O error has occurred while writing a response message entity to the container output stream.
<awoods>avmich: it sounds like they are not clustered
<avmich>I've launched rather similar config before, and got them copying data... hard to see what I made different this time
another instance complains differently:18:46
org.infinispan.commons.CacheException: Received invalid rebalance confirmation from ip-172-30-0-14-16054 for cache FedoraRepository, we don't have a rebalance in progress
maybe you're right
but why...
ERROR 22:28:50.383 (ClusteredChangeBus) Loopback changeset '8b13c2aa-be57-4dda-86aa-546b6c117269' was never received back on 'ClusteringService[cluster_name='modeshape-cluster', address=ip-172-30-0-14-22485]'. Make sure your JGroups configuration uses 'loopback=true' and if applicable 'loopback_separate_thread=true'
<awoods>indeed, but why..
* awead leaves18:48
<awoods>avmich: have you thought any more about a CLA?18:49
<avmich>ok, I'll sign :)
* cbeer leaves18:50
<awoods>avmich: only if you want to ;)
<avmich>I thought about it...
why do you feel they might be not clustered?18:54
<awoods>avmich: because they do not have replicated content18:55
<avmich>it's like by definition
if everything is done right - and yet they still for some mysterious reasons don't do what's expected - then something isn't right
<awoods>right18:56
avmich: I recently testing clustering in AWS and replication was working
s/testing/tested/
<avmich>and I don't see what I had different last time when I had them replicating... let me try another thing...
* dwilcox joins19:03
<avmich>ok, if I have clean fcrepo4-data directories in both and add a file to one, another picks it up19:06
but I want to start one with already existing data in fcrepo4-data - so that second one would replicate that19:07
and that fcrepo4-data was actually created on Fedora 4.1.x, not Fedora 4.2.0, as I'm running19:08
hmmm...
* the_mgt_ joins20:23
* the_mgt leaves20:26
* github-ff joins21:01
[fcrepo4-vagrant] mohideen opened pull request #28: Added scripts to test backups and restore of fcrepo home (master...master) http://git.io/vWbr3
* github-ff leaves
* avmich leaves21:09
* jgpawletko joins21:13
* dhlamb joins21:18
* github-ff joins21:51
[fcrepo-module-auth-webac] acoburn opened pull request #40: Implement support for acl:agentClass triples (master...fcrepo-1793) http://git.io/vWbMZ
* github-ff leaves
* jgpawletko leaves22:49
* f4jenkins leaves23:12
* f4jenkins joins
* github-ff joins23:17
[fcrepo4] acoburn opened pull request #928: define scm directly instead of relying on inheritance (master...fcrepo-1783) http://git.io/vWbx7
* github-ff leaves
* github-ff joins23:19
[fcrepo-module-auth-webac] acoburn opened pull request #41: define scm directly instead of relying on inheritance (master...fcrepo-1783) http://git.io/vWbpz
* github-ff leaves
* dhlamb leaves23:21
* github-ff joins23:22
[fcrepo4-swordserver] acoburn opened pull request #11: define scm directly instead of relying on inheritance (master...fcrepo-1783) http://git.io/vWbpF
* github-ff leaves
* dwilcox leaves00:16
* dwilcox joins00:22
* dwilcox leaves00:45
* dwilcox joins00:46
* ksclarke leaves01:15
* dwilcox leaves01:37