Log of the #duraspace-ff channel on chat.freenode.net

Using timezone: Eastern Standard Time
* kaarefc joins04:07
* kaarefc leaves04:16
* kaarefc joins04:35
* bljenkins leaves05:56
* bljenkins joins06:04
<fasseg>cbeer: eddies: im getting a lot of calls with a whopping 28% of runtime from a regexmatcher in modeshape's NameValueFactory.create(String, TextDecoder)..07:13
This gets called 2.6 million times during the creation of 338 objects
do you guys know a way to tune this or is this to deep in modeshape to change it's behaviour?
* jcoyne joins07:14
<eddies>i don't know. i think it's worth asking on the modeshape forum though07:16
or on #modeshape, but i don't know that anyone who could answer is up yet
<fasseg>yeah ill just investigate some more07:44
and ask later if I have more info07:45
eddies: this is a screenshot of my last profiling run...07:48
https://docs.google.com/file/d/0B5nd_qlYdcqyeEs5WmZMcXU1U2c/edit?usp=sharing
and I know that 800k of these calls stem from the DefaultFilter...which we could influence07:49
which is a third
Ill make the snapshot available on drive so you guys can take a look at it too..07:51
but uploading 30mb takes a while :/
and im a bit confused why RepositoryChangeBus generates ~870k events for 338 created objects07:53
<eddies>that seems…excessive07:55
<fasseg>snapshot link: https://docs.google.com/file/d/0B5nd_qlYdcqyTzF3NjlVczdORzg/edit?usp=sharing
if you check out the method list and sort by "Own Time" you'll see the matcher as the first entry07:57
which in the backtraces resolves to a call to LazyCacheNode.getChildReferences and the invocation count is split threeway there in between JcrSession.cachedNode, LazyChachedNode.parentReferenceToSelf and JcrPreSave.processAfterLocking..07:59
whith each about 870k calls
and JcrSessionCachedNode is called 870k times by the RepositoryChangeBus08:01
* JcrSession.cachedNode
In https://docs.jboss.org/author/display/MODE/Sessions+and+Listeners they suggest "If the response to a change involves much work (like processing additional content), enqueue the work and use a separate worker pool to perform the work."08:10
eddies: want me to try that ^^08:11
<eddies>fasseg: i think for the short term (i.e. before OR), we need to try things that are under our control (i.e. in org.fcrepo as opposed to MODE). So it seems looking at both DefaultFilter and the suggestion about a separate worker pool look like good candidates08:13
<fasseg>agreed, ill read up some more and investigate a bit and if I dont find any other good candidate ill try that08:15
for a single object creation I seem to get 7 NODE_ADDED events and 24 PropertyChanged events...08:42
sorry PROPERTY_ADDED events
* kaarefc leaves09:06
<pivotal-bot>Gregory Jansen added comment: "This work doesn't depend upon anything new in the BagItConnector. All the code is contained within this r..." https://www.pivotaltracker.com/story/show/5141617709:21
* gregjansen joins09:24
<pivotal-bot>Gregory Jansen added comment: "There are two commits, which maybe I should have squashed into one. In any case they are: ""09:25
https://github...." https://www.pivotaltracker.com/story/show/51415847
* github-ff joins09:26
[fcrepo4] eddies pushed 1 new commit to master: http://git.io/1PcRew
fcrepo4/master b01b335 Edwin Shin: Configures Metrics via Spring annotations.
* github-ff leaves
* github-ff joins09:27
[fcrepo-kitchen-sink] eddies pushed 1 new commit to master: http://git.io/BV_MUw
fcrepo-kitchen-sink/master 69fb903 Edwin Shin: Removes Spring xml for Metrics....
* github-ff leaves
<pivotal-bot>Gregory Jansen added comment: "I did not test by inspection for these modeshape events. I guess I could watch the events arrive at a lis..." https://www.pivotaltracker.com/story/show/5141584709:28
<bljenkins>Project fcrepo-kitchen-sink build #369: STILL UNSTABLE in 5 min 11 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/369/09:32
* travis-ci joins09:45
[travis-ci] futures/fcrepo4#678 (master - b01b335 : Edwin Shin): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/ed0af97839f6...b01b33558b32
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/8016302
* travis-ci leaves
<bljenkins>Project fcrepo-fixity-corrupter build #82: SUCCESS in 1 min 0 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/82/09:47
Project fcrepo-fixity build #263: STILL FAILING in 1 min 10 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/263/
<aawoods>has anyone else besides gregjansen built the 'fcrepo-copy-federation' project locally?09:50
<gregjansen>I doubt it. Only committed it yesterday09:51
<aawoods>I am getting pom errors
<gregjansen>oh okay.
<aawoods>and before I dig in too deeply, wanted to know if anyone else is seeing the issue.09:52
<bljenkins>Yippie, build fixed!09:54
Project fcrepo-kitchen-sink build #370: FIXED in 6 min 24 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/370/
<gregjansen>I wonder if I have things in my local maven repo that are required.. It just worked for me with "mvn -U clean verify"
I wonder if I had to built modeshape 3.4-SNAPSHOT locally09:55
<aawoods>this is what I get: https://gist.github.com/awoods/576549609:56
<gregjansen>I think I can get rid of that dependency, now that fcrepo is at 3.3 final
<aawoods>gregjansen, that would make sense.09:57
<gregjansen>in the fcrepo-copy-federation/pom.xml you can comment out the modeshape.version property near the top and then try again. that also worked for me and should pull the fcrepo specified version of mode..09:58
or I can just push that change09:59
<eddies>cbeer: joining us?10:02
<cbeer>just waiting for it to load.10:03
* nbanks_ joins10:06
* ajs6f joins
* nbanks leaves10:07
<pivotal-bot>Frank Asseg accepted "Create a simple spreadsheet for calculating required times for ingest" https://www.pivotaltracker.com/story/show/5098364510:09
Frank Asseg finished "Compare direct modeshape object creation to fcrepo4 object creation" https://www.pivotaltracker.com/story/show/51157215
Edwin Shin accepted "Compare direct modeshape object creation to fcrepo4 object creation" https://www.pivotaltracker.com/story/show/5115721510:10
Chris Beer edited "Compare direct modeshape object creation to fcrepo4 object creation" https://www.pivotaltracker.com/story/show/51157215
<ajs6f>I wish Pivotal had an "in abeyance" state.
* barmintor joins
<pivotal-bot>Edwin Shin edited "Inject authenticated sessions into the JAX-RS resources" https://www.pivotaltracker.com/story/show/4920564910:16
<aawoods>ajs6f: could you also document the use of our OAuth2, at least minimally?10:19
<pivotal-bot>A. "Yefremination" Soroka added comment: "https://github.com/futures/fcrepo-auth-oauth" https://www.pivotaltracker.com/story/show/5107764510:20
A. "Yefremination" Soroka finished "Use Apache Oltu to provide simple OAuth2 authentication" https://www.pivotaltracker.com/story/show/51077645
A. "Yefremination" Soroka accepted "Update Eclipse Formatting Rules" https://www.pivotaltracker.com/story/show/49878381
Chris Beer unstarted "make islandora work with latest build of fcrepo4. " https://www.pivotaltracker.com/story/show/48351401
<ajs6f>aawoods: Sure. A wiki page with a typical workflow?
<pivotal-bot>Chris Beer edited "Update Eclipse Formatting Rules" https://www.pivotaltracker.com/story/show/4987838110:21
Chris Beer edited "Create a simple spreadsheet for calculating required times for ingest" https://www.pivotaltracker.com/story/show/50983645
A. "Yefremination" Soroka edited "Add an OAuth2 service to the REST API's JAX-RS container using Spring Security OAuth2" https://www.pivotaltracker.com/story/show/45458249
<cbeer>ajs6f: yes. under that features, faqs and whatnot category10:22
<ajs6f>aawoods/cbber: cool.
<pivotal-bot>A. "Yefremination" Soroka added "Document our minimal OAuth impl and create tickets for follow on work" https://www.pivotaltracker.com/story/show/51565673
A. "Yefremination" Soroka started "Document our minimal OAuth impl and create tickets for follow on work" https://www.pivotaltracker.com/story/show/51565673
A. "Yefremination" Soroka edited "Document our minimal OAuth impl and create tickets for follow on work" https://www.pivotaltracker.com/story/show/5156567310:23
Chris Beer unstarted "Create a sequencer that copies projected bagit nodes to internal nodes." https://www.pivotaltracker.com/story/show/5074829110:24
Chris Beer estimated "Document the size of an AWS instance we should run performance testing against" as 2 points https://www.pivotaltracker.com/story/show/51517025
Chris Beer estimated "Run fcrepo4 single-node, async indexing configuration performance tests" as 3 points https://www.pivotaltracker.com/story/show/51516221
Chris Beer estimated "Run fcrepo4 single-node configuration performance tests" as 3 points https://www.pivotaltracker.com/story/show/51516137
Chris Beer estimated "Run fcrepo3 performance tests on AWS" as 3 points https://www.pivotaltracker.com/story/show/51515987
Chris Beer estimated "Create asynchronous indexing, clustered configuration for modeshape" as 3 points https://www.pivotaltracker.com/story/show/51515429
Chris Beer estimated "Investigate alternative indexStorage options (especially in a clustered configuration)" as 2 points https://www.pivotaltracker.com/story/show/51417691
Chris Beer estimated "Use good predicates and document them in futures/ontology" as 3 points https://www.pivotaltracker.com/story/show/50895969
Chris Beer estimated "Add enunciate docs to the webapp" as 2 points https://www.pivotaltracker.com/story/show/5082721910:25
Chris Beer estimated "(stub) Sonar hotspot cleanup ticket" as 5 points https://www.pivotaltracker.com/story/show/51518237
Chris Beer estimated "(stub) Spring Config Cleanup" as 3 points https://www.pivotaltracker.com/story/show/51518197
Chris Beer added "Add test coverage for JcrRdfTools" https://www.pivotaltracker.com/story/show/51565867
Chris Beer edited "Add test coverage for JcrRdfTools" https://www.pivotaltracker.com/story/show/51565867
Chris Beer added "Add test coverage to FedoraResource" https://www.pivotaltracker.com/story/show/5156594310:26
Chris Beer edited "Add test coverage to FedoraResource" https://www.pivotaltracker.com/story/show/51565943
<cbeer>ajs6f: in fact, i have just the page for you: https://wiki.duraspace.org/display/FF/Authentication+and+Authorization10:27
<pivotal-bot>Chris Beer edited "Document our minimal OAuth impl and create tickets for follow on work" https://www.pivotaltracker.com/story/show/51565673
<ajs6f>cbeer: Thanks, I was a bout to start a new one.
<pivotal-bot>Chris Beer added "Document process for getting fcrepo4, running it, and getting into the REST API" https://www.pivotaltracker.com/story/show/5156607110:28
Chris Beer edited "Document process for getting fcrepo4, running it, and getting into the REST API" https://www.pivotaltracker.com/story/show/51566071
Chris Beer added "Document our use of sparql update to update object properties" https://www.pivotaltracker.com/story/show/51566103
Chris Beer edited "Document our use of sparql update to update object properties" https://www.pivotaltracker.com/story/show/51566103
Edwin Shin estimated "Document the size of an AWS instance we should run performance testing against" as 1 point https://www.pivotaltracker.com/story/show/5151702510:29
Chris Beer added comment: "https://wiki.duraspace.org/display/FF/Performance+Testing" https://www.pivotaltracker.com/story/show/51517025
Chris Beer edited "Run fcrepo4 single-node, async indexing configuration performance tests" https://www.pivotaltracker.com/story/show/51516221
Chris Beer edited "Run fcrepo3 performance tests on AWS" https://www.pivotaltracker.com/story/show/51515987
Edwin Shin edited "Run fcrepo4 single-node configuration performance tests" https://www.pivotaltracker.com/story/show/51516137
Andrew Woods edited "Document the size of an AWS instance we should run performance testing against" https://www.pivotaltracker.com/story/show/51517025
Chris Beer edited "Create asynchronous indexing, clustered configuration for modeshape" https://www.pivotaltracker.com/story/show/51515429
Chris Beer edited "Improve hydra integration test response time" https://www.pivotaltracker.com/story/show/51309207
<cbeer>aawoods: https://github.com/infinispan/infinispan/blob/master/core/src/main/resources/jgroups-ec2.xml10:32
<pivotal-bot>Chris Beer edited "Run fcrepo4 single-node configuration performance tests" https://www.pivotaltracker.com/story/show/5151613710:33
Edwin Shin edited "Run fcrepo4 single-node, async indexing configuration performance tests" https://www.pivotaltracker.com/story/show/5151622110:34
Edwin Shin edited "Run fcrepo4 single-node configuration performance tests" https://www.pivotaltracker.com/story/show/5151613710:35
Chris Beer added "Add unit test coverage to FedoraVersions" https://www.pivotaltracker.com/story/show/5156671310:36
Chris Beer edited "Add unit test coverage to FedoraVersions" https://www.pivotaltracker.com/story/show/51566713
<ajs6f>Is that our benchmark?10:37
<pivotal-bot>Chris Beer added "Add unit test coverage to uri resource graph injection classes" https://www.pivotaltracker.com/story/show/51566849
Chris Beer edited "Add unit test coverage to uri resource graph injection classes" https://www.pivotaltracker.com/story/show/51566849
Edwin Shin edited "Publish JMeter reports to Sonar via Maven" https://www.pivotaltracker.com/story/show/5151775310:38
Andrew Woods edited "Publish JMeter reports to Sonar via Maven" https://www.pivotaltracker.com/story/show/51517753
Edwin Shin added "Document our AWS benchmarking setup" https://www.pivotaltracker.com/story/show/5156703310:40
Edwin Shin edited "Document our AWS benchmarking setup" https://www.pivotaltracker.com/story/show/51567033
Edwin Shin edited "Document our AWS benchmarking setup" https://www.pivotaltracker.com/story/show/51567033
Edwin Shin edited "hydra + fcrepo4 benchmark" https://www.pivotaltracker.com/story/show/5068293710:41
Edwin Shin edited "hydra + fcrepo4 benchmark" https://www.pivotaltracker.com/story/show/50682937
Chris Beer added "Add REST API documentation about WHY we made the decisions we did" https://www.pivotaltracker.com/story/show/51567165
Chris Beer edited "Add REST API documentation about WHY we made the decisions we did" https://www.pivotaltracker.com/story/show/51567165
Chris Beer estimated "Document process for getting fcrepo4, running it, and getting into the REST API" as 1 point https://www.pivotaltracker.com/story/show/5156607110:42
Andrew Woods added "Enable auto-clustering of servers deployed with Puppet" https://www.pivotaltracker.com/story/show/51567187
Andrew Woods estimated "Enable auto-clustering of servers deployed with Puppet" as 0 points https://www.pivotaltracker.com/story/show/51567187
Andrew Woods edited "Enable auto-clustering of servers deployed with Puppet" https://www.pivotaltracker.com/story/show/51567187
Edwin Shin edited "Run fcrepo3 performance tests on AWS" https://www.pivotaltracker.com/story/show/51515987
Edwin Shin added "Generate AWS IAM accounts for Sprint 12 devs" https://www.pivotaltracker.com/story/show/5156737110:44
Andrew Woods added "Update Cloud Tooling to support Provisioned IOPS" https://www.pivotaltracker.com/story/show/51567391
Andrew Woods edited "Update Cloud Tooling to support Provisioned IOPS" https://www.pivotaltracker.com/story/show/5156739110:45
Andrew Woods estimated "Update Cloud Tooling to support Provisioned IOPS" as 2 points https://www.pivotaltracker.com/story/show/51567391
Edwin Shin edited "Generate AWS IAM accounts for Frank" https://www.pivotaltracker.com/story/show/5156737110:46
Edwin Shin edited "Generate AWS IAM accounts for Frank" https://www.pivotaltracker.com/story/show/51567371
<cbeer>ajs6f: can you explain https://www.pivotaltracker.com/story/show/51518197?
<pivotal-bot>feature: (stub) Spring Config Cleanup (unstarted) / owner:
<cbeer>we weren't sure what the discrete tasks was
<pivotal-bot>Andrew Woods edited "Enable auto-clustering of servers deployed with Puppet" https://www.pivotaltracker.com/story/show/5156718710:47
Edwin Shin edited "Enable auto-clustering of servers deployed with Puppet" https://www.pivotaltracker.com/story/show/51567187
Edwin Shin edited "Improve hydra integration test response time" https://www.pivotaltracker.com/story/show/51309207
Andrew Woods estimated "Enable auto-clustering of servers deployed with Puppet" as 1 point https://www.pivotaltracker.com/story/show/5156718710:48
<ajs6f>cbeer: No real idea, at this point. I think we were talking about how there's too much Spring XML floating around and it's too redundant. I think maybe eddies suggested using programmatic config. We also talked about having a "wiring" module from which Spring XML could be imported.
<cbeer>ajs6f: so.. what's the strategy for limiting the amount of xml floating around? generating test-jars?10:49
<ajs6f>cbeer: That's what we didn't decide. We just decided that it sucks to have it all floating around.
<cbeer>ajs6f: so. how do we decide?
<ajs6f>cbeer: Well, if we stay committed to Spring (and I think we are for the moment) we could engage more fully with scanning. Make more things components. That would reduce the XML sharply.10:50
<pivotal-bot>Edwin Shin added "Analyze single-node sync & async test results and generate new tickets" https://www.pivotaltracker.com/story/show/51567801
<ajs6f>cbeer: Or we could go to programmatic config and build up a config type system.
<cbeer>ajs6f: so, the services?
i think those are the big offenders, right?
<ajs6f>cbeer: They're part of it.
<pivotal-bot>Edwin Shin edited "Analyze single-node sync & async test results and generate new tickets" https://www.pivotaltracker.com/story/show/51567801
Frank Asseg added "Investigate the created profiling results " https://www.pivotaltracker.com/story/show/51567861
<ajs6f>cbeer: I'd like to take the CDI stance on this (not actually using CDI).10:51
<pivotal-bot>Frank Asseg edited "Investigate the created profiling results " https://www.pivotaltracker.com/story/show/51567861
Andrew Woods added "Enable remote YourKit profiling with Puppet" https://www.pivotaltracker.com/story/show/51567901
<ajs6f>cbeer; meaning that the application should autowire as much as possible and explicit config should only discuss choices an integrator might need or want to make.
<pivotal-bot>Edwin Shin added "Analyze clustered test results and generate new tickets for improving performance" https://www.pivotaltracker.com/story/show/51567907
Andrew Woods edited "Enable remote YourKit profiling with Puppet" https://www.pivotaltracker.com/story/show/51567901
Edwin Shin edited "Analyze clustered test results and generate new tickets for improving performance" https://www.pivotaltracker.com/story/show/51567907
<ajs6f>cbeer: e.g. for the services,
we know that we need them. They aren't optional.
<cbeer>ajs6f: ok. so, what are the choices we want to expose? i guess the modeshape config?10:52
<ajs6f>There's no reason to explcitly wire them all the time.
cbeer; Yep. And the serializers.
<cbeer>ajs6f: although the serializers could be.. if they're present, use them.
<ajs6f>cbeer: And JMS stuff.
<cbeer>oh, yeah. jms.
<ajs6f>cbeer: yep. That would be better.
<pivotal-bot>Edwin Shin edited "Enable remote YourKit profiling with Puppet" https://www.pivotaltracker.com/story/show/51567901
<cbeer>ajs6f: so scan the services.10:53
and scan the serializers
<ajs6f>cbeer: yep.
SCAN ALL THE THINGS!
<cbeer>(i think i did something hacky with the uri resource-injecting stuff that could be reused)
<ajs6f>scan the session factory and provider.
<pivotal-bot>Frank Asseg edited "Investigate the created profiling results " https://www.pivotaltracker.com/story/show/5156786110:54
Edwin Shin deleted "(stub) Sonar hotspot cleanup ticket" https://www.pivotaltracker.com/story/show/51518237
Chris Beer added "Use component scanning to inject serializers" https://www.pivotaltracker.com/story/show/51568127
Chris Beer edited "Use component scanning to inject serializers" https://www.pivotaltracker.com/story/show/51568127
Chris Beer added "Use component scanning to inject required services" https://www.pivotaltracker.com/story/show/5156814310:55
Chris Beer edited "Use component scanning to inject required services" https://www.pivotaltracker.com/story/show/51568143
* ajs6f leaves10:57
<pivotal-bot>Edwin Shin edited "Create new, specific performance enhancement tickets based on the profiling investigation" https://www.pivotaltracker.com/story/show/5156786110:58
Chris Beer added "Improve performance of SPARQL/UPDATE request endpoint to take <150ms" https://www.pivotaltracker.com/story/show/5156849710:59
Chris Beer edited "Improve performance of SPARQL/UPDATE request endpoint to take <150ms" https://www.pivotaltracker.com/story/show/5156849711:00
Edwin Shin edited "Create new, specific performance enhancement tickets based on the profiling investigation" https://www.pivotaltracker.com/story/show/51567861
<aawoods>I will retry testing the bagit-federation when you push your update, gregjansen.
* ajs6f joins11:03
<pivotal-bot>Chris Beer edited "Create new, specific performance enhancement tickets based on the profiling investigation" https://www.pivotaltracker.com/story/show/5156786111:04
<gregjansen>aawoods: pushed11:05
<aawoods>gregjansen: you da man
it even builds, which is a treat.11:06
<pivotal-bot>Chris Beer added "Document the list of "features" in the alpha" https://www.pivotaltracker.com/story/show/5156905511:07
Chris Beer edited "Document the list of "features" in the alpha" https://www.pivotaltracker.com/story/show/51569055
Chris Beer added "Document recommended heap and stack sizes for running fcrepo4" https://www.pivotaltracker.com/story/show/5156913911:08
Chris Beer edited "Document recommended heap and stack sizes for running fcrepo4" https://www.pivotaltracker.com/story/show/51569139
Chris Beer added comment: "still broken. " https://www.pivotaltracker.com/story/show/46108229
Chris Beer unstarted "Backup a running repository" https://www.pivotaltracker.com/story/show/4610822911:09
Chris Beer deleted ""identity" sequencer" https://www.pivotaltracker.com/story/show/5074797911:10
Edwin Shin deleted "(stub) Spring Config Cleanup" https://www.pivotaltracker.com/story/show/5151819711:13
Chris Beer estimated "Use component scanning to inject serializers" as 3 points https://www.pivotaltracker.com/story/show/5156812711:14
Chris Beer estimated "Use component scanning to inject required services" as 1 point https://www.pivotaltracker.com/story/show/51568143
Chris Beer estimated "Document the list of "features" in the alpha" as 2 points https://www.pivotaltracker.com/story/show/51569055
Chris Beer added "Tag a 4.0.0-ALPHA1 release of fcrepo4 and publish the artifacts" https://www.pivotaltracker.com/story/show/5156988111:15
A. "Yefremination" Soroka added "Protect the /authorize endpoint in the OAuth way" https://www.pivotaltracker.com/story/show/51569887
A. "Yefremination" Soroka estimated "Protect the /authorize endpoint in the OAuth way" as 3 points https://www.pivotaltracker.com/story/show/51569887
Chris Beer edited "Tag a 4.0.0-ALPHA1 release of fcrepo4 and publish the artifacts" https://www.pivotaltracker.com/story/show/51569881
A. "Yefremination" Soroka added comment: "https://www.pivotaltracker.com/story/show/51569887" https://www.pivotaltracker.com/story/show/5156567311:16
A. "Torbulater" Soroka added "Remove OAuth workflows that we don't want to support" https://www.pivotaltracker.com/story/show/5157012311:18
Chris Beer added "Create a jetty assembly package for the webapp" https://www.pivotaltracker.com/story/show/51570127
Chris Beer edited "Create a jetty assembly package for the webapp" https://www.pivotaltracker.com/story/show/51570127
A. "Torbulater" Soroka added comment: "https://www.pivotaltracker.com/story/show/51570123" https://www.pivotaltracker.com/story/show/51565673
Chris Beer edited "Document process for getting fcrepo4, running it, and getting into the REST API" https://www.pivotaltracker.com/story/show/5156607111:19
Andrew Woods added comment: "It would also be helpful to include "how to use our OAuth" in the documentation." https://www.pivotaltracker.com/story/show/5156567311:20
A. "Torbulater" Soroka added comment: "I'm doing that on the wiki page. You can look at the draft now. I'm going to add a few example UR..." https://www.pivotaltracker.com/story/show/5156567311:21
Chris Beer edited "Create a hydra-jetty fcrepo4 fork" https://www.pivotaltracker.com/story/show/51570127
Chris Beer edited "Add REST API documentation about WHY we made the decisions we did" https://www.pivotaltracker.com/story/show/5156716511:24
Chris Beer edited "Document our use of sparql update to update object properties" https://www.pivotaltracker.com/story/show/51566103
Edwin Shin edited "Document process for getting fcrepo4, running it, and getting into the REST API" https://www.pivotaltracker.com/story/show/5156607111:28
Edwin Shin edited "Document the list of "features" in the alpha" https://www.pivotaltracker.com/story/show/51569055
* gregjansen leaves
<pivotal-bot>Edwin Shin edited "Add REST API documentation about WHY we made the decisions we did" https://www.pivotaltracker.com/story/show/5156716511:29
Edwin Shin edited "Document our use of sparql update to update object properties" https://www.pivotaltracker.com/story/show/51566103
Andrew Woods edited "Document the list of "features" in the alpha" https://www.pivotaltracker.com/story/show/51569055
Edwin Shin edited "Document recommended heap and stack sizes for running fcrepo4" https://www.pivotaltracker.com/story/show/51569139
A. "Torbulater" Soroka added "Stop recreating a test namespace every time a JAX-RS resource is created." https://www.pivotaltracker.com/story/show/5157117311:30
Edwin Shin edited "Write documentation for running fcrepo 4.x as a WAR under e.g. tomcat" https://www.pivotaltracker.com/story/show/45507631
Edwin Shin deleted "Write documentation for running fcrepo 4.x as a WAR under e.g. tomcat" https://www.pivotaltracker.com/story/show/45507631
* gregjansen joins11:31
* gregjansen leaves
<ajs6f>barmintor: re: the injected Session thing… did we ever figure out whether it is even possible to do field injection with that Jersey provider stuff? Did we end up thinking it's limited to parameter injection11:34
?
<barmintor>ajs6f: session depends on request, so it can be injected as a parm
<cbeer>eddies: https://wiki.duraspace.org/display/FF/Hydra+and+Fedora+4
<ajs6f>barmintor: yeah, that's cool. But can we do it as a field, or do we have to clutter the method sigs?11:35
<barmintor>I don't think you can do request-scoped field injection.
<ajs6f>D'OH!
<barmintor>Try it and see, though11:36
<ajs6f>I just did. It's not clear whether it works.
The tests fail inconclusively.
They run until the mthod tries to log out the session.
<pivotal-bot>Chris Beer added "Document the fcrepo4 "object model"" https://www.pivotaltracker.com/story/show/5157176711:37
Chris Beer edited "Document the fcrepo4 "object model"" https://www.pivotaltracker.com/story/show/51571767
<cbeer>eddies: oh, i shouldn't have done a hangout-on-air and recorded it to youtube?11:41
<pivotal-bot>Chris Beer added "Stub a migration document that says "coming soon"" https://www.pivotaltracker.com/story/show/5157213111:42
Chris Beer edited "Stub a migration document that says "coming soon"" https://www.pivotaltracker.com/story/show/51572131
A. "Torbulater" Soroka edited "Document the fcrepo4 "object model"" https://www.pivotaltracker.com/story/show/51571767
Chris Beer edited "Stub a migration document that says "coming soon"" https://www.pivotaltracker.com/story/show/5157213111:44
Edwin Shin added "Produce a set of performance charts comparing fcrepo3 and fcrepo4" https://www.pivotaltracker.com/story/show/51572283
Edwin Shin edited "Produce a set of performance charts comparing fcrepo3 and fcrepo4" https://www.pivotaltracker.com/story/show/51572283
<ajs6f>barmintor: oh, fooey. See:
https://java.net/jira/browse/JERSEY-513
If James Strachan can't do it, I can't.
<barmintor>Is that actually what we're talking aboiut, though?11:45
<ajs6f>It's not _exactly_ the same thing, but if they can't inject their own @Context stuff, they can't do ours, right?11:46
<barmintor>But they do inject some context stuff
<ajs6f>Yeah, I know.
<pivotal-bot>Chris Beer estimated "Create new, specific performance enhancement tickets based on the profiling investigation" as 3 points https://www.pivotaltracker.com/story/show/5156786111:47
Andrew Woods deleted "Establish performance baseline against cluster configuration" https://www.pivotaltracker.com/story/show/5115190311:48
Andrew Woods edited "Update Cloud Tooling to support Provisioned IOPS" https://www.pivotaltracker.com/story/show/51567391
Andrew Woods edited "Enable remote YourKit profiling with Puppet" https://www.pivotaltracker.com/story/show/51567901
Gregory Jansen added comment: "I wrote the query code and added it to the ManifestMonitorIT. However it is commented out b/c it doesn't ..." https://www.pivotaltracker.com/story/show/5074867711:59
Chris Beer started "Use component scanning to inject serializers" https://www.pivotaltracker.com/story/show/51568127
Chris Beer started "Use component scanning to inject required services" https://www.pivotaltracker.com/story/show/51568143
<ajs6f>Holy *&)*&. I think it worked.12:14
That's certainly unexpected behavior.
Oh, wait, no.12:15
Good, that's what I wanted to see. Utter failure.
<pivotal-bot>Edwin Shin added comment: "https://wiki.duraspace.org/display/FF/Test+Platform" https://www.pivotaltracker.com/story/show/5151702512:23
Edwin Shin finished "Document the size of an AWS instance we should run performance testing against" https://www.pivotaltracker.com/story/show/51517025
<cbeer>eddies: can you add how to get an AWS system like that? or is it blatently obvious if i bothered to go looking?12:24
<eddies>i don't know how it's done via Puppet, which is I think what we're aiming for12:25
via the web console, it's just point and click
<aawoods>cbeer: with the puppet setup, creating an m1.large is trivial12:26
<cbeer>ok, sounds good. i guess aawoods must have a good idea about that already
<aawoods>there is another ticket to get the ebs-optimized and provisioned iops integrated into the tooling.12:27
cbeer: I would be happy to discuss... and I will be documenting the process.
<cbeer>aawoods: that's all i'm looking for, thanks. provisioned iops is new to me12:28
<eddies>aawoods: i don't know if you're already doing this, but until you get to testing puppet w/ ebs-optimized + iiops, i'd get everything else sorted out with a micro instance to keep our costs down =)12:30
<aawoods>eddies: right12:31
<pivotal-bot>Chris Beer finished "Use component scanning to inject required services" https://www.pivotaltracker.com/story/show/5156814312:40
Chris Beer added comment: "https://github.com/futures/fcrepo4/commit/2514d1bdee70d304eacde44b998546362680ae77" https://www.pivotaltracker.com/story/show/5156814312:41
* github-ff joins
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/UFI_0A
fcrepo4/master 2514d1b Chris Beer: use component scanning for discovering required services
* github-ff leaves
<cbeer>ajs6f: so, i'm looking at the serializers now12:43
i think we need to add an annotation or class-level method for the "key"
oh, i guess it isn't a class-level method12:44
<pivotal-bot>Andrew Woods started "Update Cloud Tooling to support Provisioned IOPS" https://www.pivotaltracker.com/story/show/5156739112:45
* ff-github joins12:46
[fcrepo-fixity] fasseg pushed 4 new commits to master: http://git.io/rQORiQ
fcrepo-fixity/master 6d0175f fasseg: added versions to pom.xml that has been removed from the parent; add test resources
fcrepo-fixity/master 1a892a2 fasseg: updated ITs to use the new ContainerWrapper approach
fcrepo-fixity/master 3965d86 fasseg: fixed broken test:FedoraFixityClientIT. testRetrievePids()
* ff-github leaves
<bljenkins>Project fcrepo-fixity build #264: NOW UNSTABLE in 7 min 47 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/264/12:54
* travis-ci joins13:00
[travis-ci] futures/fcrepo4#679 (master - 2514d1b : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/b01b33558b32...2514d1bdee70
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/8023420
* travis-ci leaves
<ajs6f>cbeer: What's up with the serializers? I don't quite understand the question...13:09
<cbeer>ajs6f: after the tests finish, i'll send a PR and you can see what i've done13:10
<bljenkins>Project fcrepo-fixity-corrupter build #83: SUCCESS in 4 min 36 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/83/
<ajs6f>ok
<bljenkins>Yippie, build fixed!13:18
Project fcrepo-fixity build #265: FIXED in 12 min: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/265/
* nbanks_ joins13:20
* nbanks leaves13:21
<bljenkins>Project fcrepo-kitchen-sink build #371: SUCCESS in 6 min 11 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/371/
* github-ff joins13:29
[fcrepo4] cbeer created serialization-autodiscovery (+1 new commit): http://git.io/z9FAJA
fcrepo4/serialization-autodiscovery b2a793f Chris Beer: auto-discover serialization implementations, allowing the serializers to be scanned and injected without additional spring configuration
* github-ff leaves
<cbeer>ajs6f: https://github.com/futures/fcrepo4/pull/78
* github-ff joins
[fcrepo4] cbeer opened pull request #78: auto-discover serialization implementations, allowing the serializers to... (master...serialization-autodiscovery) http://git.io/CdLDOg
* github-ff leaves
<ajs6f>Reviewing it now...13:32
M, I'll wait for travis, just in case
<pivotal-bot>Chris Beer finished "Use component scanning to inject serializers" https://www.pivotaltracker.com/story/show/5156812713:36
Chris Beer added comment: " http://git.io/CdLDOg" https://www.pivotaltracker.com/story/show/5156812713:37
Chris Beer edited "Use component scanning to inject serializers" https://www.pivotaltracker.com/story/show/51568127
<cbeer>i think the rest of the spring wiring is for stuff that needs configuration
or we probably want a choice about impls
<ajs6f>Well, that's fair game for explicit config, in my book.13:38
Okay, got the injected field thing to work.13:48
What I bloody nightmare.
<bljenkins>Project fcrepo-fixity-corrupter build #84: SUCCESS in 56 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/84/
<ajs6f>s/I/a
Okay, maybe I'm a bloody nightmare, too.
I had to move the provider out of the remit of Spring into Jersey's DI, but still let Spring supply the dependencies for the provider (like a SessionFactory and Repo and so forth).13:49
eddies: per discussion this morning, I'll move on to editing all the API methods to use this.
<bljenkins>Project fcrepo-fixity build #266: SUCCESS in 4 min 56 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/266/13:52
<cbeer>yourkit++13:53
<ajs6f>It's pretty sweet. It's the only profiler I've used in years, tho', so I do wonder what else is out there. That's just how I am. I'm a player.
<fasseg>futures6's tomcat7 produces this error message when starting fcrepo4: There is no namespace registered for the prefix "webhook" when Spring tries to create the modeshaperRepoFactory bean...Has anyone seen this before?13:54
<ajs6f>Are we putting webhooks in a namespace? Wouldn't we rather put them in a workspace?
<fasseg>this is the full error message: Error creating bean with name 'modeshapeRepofactory': Invocation of init method failed; nested exception is javax.jcr.RepositoryException: org.modeshape.jcr.value.ValueFormatException: Error converting "webhook:resource" from String to a Name
<bljenkins>Project fcrepo-kitchen-sink build #372: SUCCESS in 6 min 14 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/372/13:55
<cbeer>fasseg: i renamed the webhooks namespace yesterday.. probably need to remove the jcr data directory and restart13:59
i can do that after these tests finish
<ajs6f>cbeer: what's the diff between FedoraRepositoryImport and FedoraImport?
<cbeer>ajs6f: FedoraRepositoryImport is mounted at the root level.
<fasseg>ahh I see that's why I cant reproduce this locally
<ajs6f>cbeer: Thanks— so that's all about JAX-RS pathing?
<cbeer>yep. not sure if you're looking at my PR, but i made FedoraRepositoryImport extend FedoraImport there14:00
<fasseg>cbeer: that's "/opt/fcrepo4-data/" on futures6?
<cbeer>fasseg: yes. leave the directory, but delete its children.14:01
<fasseg>kk thanks
* nbanks_ leaves14:06
<ajs6f>cbeer: yeah, that's what weirded me out— I didn't look at the @Path.14:07
anyways, sounds good.
cbeer: so the map of serliazlizers get built because the app context gets injected via mutator, so that the SerializerUtil can retrieve them?14:09
cause the only thing I see is that maybe we want to build the map in a an initialzier instead of at request time.
That might avoid some uneccesary calls into the app context.14:10
<cbeer>ajs6f: i thought about that. i wasn't sure what point in the life-cycle that class would get built
* nbanks joins14:11
<ajs6f>It's a plain Jane @Component, so once, at startup, before it gets injected into the JAX-RS. We could do a @PostConstruct public void buildMySerializerMap() {}
<cbeer>k
<ajs6f>All righty then.14:13
<cbeer>ok, running tests then i'll merge.14:16
* github-ff joins14:17
[fcrepo4] ajs6f created NewInjectedSessions (+1 new commit): http://git.io/5MUCfQ
fcrepo4/NewInjectedSessions ee0cd2b ajs6f: Field injection for JCR Sessions
* github-ff leaves
<ajs6f>Excellente.
* github-ff joins14:20
[fcrepo4] cbeer merged serialization-autodiscovery into master: http://git.io/GcN1Tw
* github-ff leaves
* bljenkins leaves14:24
<ajs6f> We killed the RDF generator. Do we want to eventually kill the DC generator? After all, you can just put your DC assertions into properties....
* bljenkins joins14:28
<cbeer>ajs6f: that's less clear to me.. but i'm not sure anyone would be sad to see it go.14:31
<ajs6f>Oh, I know we'd all _like_ to kill it. I guess I should ask Anusha. She made the loudest arguments for it.
* github-ff joins14:37
[fcrepo4] ajs6f pushed 2 new commits to NewInjectedSessions: http://git.io/ZvDmQw14:38
fcrepo4/NewInjectedSessions f5435f8 ajs6f: Injecting FedoraDatastreams and correcting fcrepo-generator-dc
fcrepo4/NewInjectedSessions 0f56e1f ajs6f: Done through FedoraExport
* github-ff leaves
<bljenkins>Project fcrepo-fixity-corrupter build #85: SUCCESS in 41 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/85/
<ajs6f>afk bbi1014:40
* ajs6f leaves
<bljenkins>Project fcrepo-fixity build #267: SUCCESS in 5 min 13 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/267/14:42
* ajs6f joins14:43
* travis-ci joins14:46
[travis-ci] futures/fcrepo4#681 (master - a35afee : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/2514d1bdee70...a35afeed07a5
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/8026232
* travis-ci leaves
* bljenkins leaves14:49
* bljenkins joins14:51
* github-ff joins
[fcrepo4] ajs6f pushed 1 new commit to NewInjectedSessions: http://git.io/Yxz93A
fcrepo4/NewInjectedSessions e602ba7 ajs6f: Done through FedoraFixity
* github-ff leaves
<bljenkins>Project fcrepo-fixity-corrupter build #86: SUCCESS in 55 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/86/14:57
* nbanks leaves15:00
* nbanks joins
<bljenkins>Project fcrepo-fixity build #268: UNSTABLE in 3 min 56 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/268/
* nbanks leaves15:05
* nbanks joins15:06
<bljenkins>Project fcrepo4 build #798: UNSTABLE in 13 min: http://ci.projectblacklight.org/jenkins/job/fcrepo4/798/15:10
<ajs6f>fasseg: I just changed the FedoraTrtansactions endpoint to use injected Sessions.15:13
Now the i-test fails, because apparently transactions never get reaped.
Any idea why?
<fasseg>the reaper is using the Scheduled annotation from spring, is the spring task scheduler bean still created?15:14
<ajs6f>I didn't touch that at all.
I did change the endpoint bean to be scope-prototype, because that's the only way injection can work.15:15
Could that be the problem?
<cbeer>hm. this is promising... i created 100 objects and a file datastream (with a short random payload) in a hydra head
and the timing for fcrepo4 is on par with fcrepo3 (faster, but probably within a margin of error.)
<fasseg>ajs6f: yeah, since the @Scheduled annotation is inside the endpoint bean, this might be the cause15:16
<ajs6f>Well, that's a problem.
It makes no sense to have field injection for a request-scoped object in a singleton bean.
Maybe we can move the reaper out of the endpoint?
<fasseg>just move?15:17
<ajs6f>Yeah, have it live in a different bean, a singletone bean.
<fasseg>and use it as a service in the enpoint?
<ajs6f>Why would the endpoint need it?
<fasseg>hmm currently transations are transiently kept in a hsashmap in FedoraTransactions15:18
<pivotal-bot>Chris Beer started "hydra + fcrepo4 benchmark" https://www.pivotaltracker.com/story/show/50682937
<fasseg>and the reaper accesses this hashmap and removes elements if neccessary
<pivotal-bot>Chris Beer delivered "Use component scanning to inject required services" https://www.pivotaltracker.com/story/show/51568143
Chris Beer delivered "Use component scanning to inject serializers" https://www.pivotaltracker.com/story/show/51568127
<ajs6f>How about I do all the other endpoints, and you can take a look at that.
<fasseg>this was just a strawman impl i remind you ;)
<ajs6f>We burn strawmen for fun.
<fasseg>cant today, gonna be tomorrow then15:19
<ajs6f>Okay,
.
<fasseg>ajs6f: but injection of services is possible right? so I could create a TransactionService for wrapping that logic and it would work with your patch?15:20
<ajs6f>Yeeeahhh, I think so. I still don't understand why the endpoint should care about the reaper. The endpoint just needs to have access to a store of transactions.15:21
And so does the reaper.
<fasseg>not the reaper, but a TransactionService which would keep transactions in a datastructure, and conviently this also has the reaper method15:22
so the endpoint would delegate to the TransactionService for create/rollback
<ajs6f>It was conveniently, but perhaps no longer so convenient, if it causes a scope mistmatch.
Yeah, I think a T-service is the way to go.15:23
That's been a workable pattern so far.
<fasseg>that's what I'm saying
and this service has the reaper method as well
<ajs6f>Okay. I'll finish the other endpoints and you can take a shot at the t-service when you can.
<fasseg>not the enpoint
<ajs6f>Yeah, yeah. I get it. I think. The important thing for me is to get the reaper and store of tsactions out of the endpoint. That's just good JAX-RS practice, anyway. Endpoints should be almost pure controllers.15:25
* github-ff joins15:26
[fcrepo4] ajs6f pushed 1 new commit to NewInjectedSessions: http://git.io/NFrGIA
fcrepo4/NewInjectedSessions 56247d8 ajs6f: All endpoints except for FedoraTransactions done.
* github-ff leaves
<cbeer>aawoods: FF phase 2 is through dec 2013, right?15:31
<aawoods>yes, although I believe it is technically called phase 3
I call it beta
<cbeer>ok. beta through dec 2013.
<aawoods>or beta-alpha
<cbeer>and 4.0 through.. june 2014?
<aawoods>sounds right for new15:32
<cbeer>k
<aawoods>now
* ff-github joins15:43
[fcrepo-fixity] fasseg pushed 2 new commits to master: http://git.io/j4CFLA
fcrepo-fixity/master 324eb4e fasseg: removed uneccessary log statement which sonar complained about
fcrepo-fixity/master 4c8a65f fasseg: fixed states to use the available states in fcrepo's sources. Fixed Javascript to show a Piechart when only errors are in the database
* ff-github leaves
<cbeer>i probably should have realized that a test with 100 objects that takes 3 minutes
when scaled to 10000 objects will take 300 minutes.
<ajs6f>Depends on Amdahl's law. :)15:44
<cbeer>not sure how much that applies to fcrepo 315:45
* ajs6f leaves15:47
<bljenkins>Project fcrepo-fixity build #269: STILL UNSTABLE in 5 min 40 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/269/15:48
* nbanks leaves15:51
* ajs6f joins15:53
<bljenkins>Yippie, build fixed!15:54
Project fcrepo4 build #799: FIXED in 25 min: http://ci.projectblacklight.org/jenkins/job/fcrepo4/799/
Project fcrepo-fixity-corrupter build #87: SUCCESS in 1 min 14 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/87/15:55
Project fcrepo-fixity build #270: STILL UNSTABLE in 7 min 6 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/270/16:01
Project fcrepo-kitchen-sink build #373: UNSTABLE in 6 min 31 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/373/16:02
<ajs6f>cbeer: Got a minute?16:10
* github-ff joins
[fcrepo4] ajs6f force-pushed NewInjectedSessions from 56247d8 to 047aaf2: http://git.io/pBY7mQ
fcrepo4/NewInjectedSessions c5138c4 ajs6f: Field injection for JCR Sessions
fcrepo4/NewInjectedSessions e3f81c9 ajs6f: Injecting FedoraDatastreams and correcting fcrepo-generator-dc
fcrepo4/NewInjectedSessions d41a6a7 ajs6f: Done through FedoraExport
* github-ff leaves
<cbeer>ajs6f: shoot.
<ajs6f>I'm almost through the injected-sessions thing.16:11
There's an issue with Transactions, but fasseg understands it and has a solution in mind.
But when I rebased over your recent changes to serializtion...
svery oddly, the FedoraImportTest blew up, and I can't for the life of me see why.16:12
It's throwing an NPE out of the session.logout()
(in the actual JAX-RS method)
a few lines after it happily does session.save()!
I think I must be missing something obvious.16:13
* nbanks joins
<cbeer>huh. is that in NewInjectedSessions?
<ajs6f>Yup.
Just pushed the curent state.
<cbeer>and just FedoraImportTest? or also FedoraRepositoryImportTest?16:14
<ajs6f>I don't see any FedoraRepositoryImportTest.
Should I? That's odd.
No, I don't see any such guy in master, either.16:15
<cbeer>https://github.com/futures/fcrepo4/blob/NewInjectedSessions/fcrepo-http-api/src/test/java/org/fcrepo/api/repository/FedoraRepositoryImportTest.java
<ajs6f>This is weird. I don't see that guy here:16:16
https://github.com/futures/fcrepo4/tree/master/fcrepo-http-api/src/test/java/org/fcrepo/api
and I don't see him in my local repo.
(which i rebased over master after your merge)16:17
OH! OH!
You're right.
I was misreading the stack trace.
What an idiot!
back to work.
* nbanks leaves16:21
<bljenkins>Project fcrepo-fixity-corrupter build #88: SUCCESS in 35 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/88/16:26
Project fcrepo-fixity build #271: STILL UNSTABLE in 2 min 36 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/271/16:28
Project fcrepo-kitchen-sink build #374: STILL UNSTABLE in 2 min 20 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/374/
<pivotal-bot>A. "Torbulater" Soroka added comment: "https://github.com/futures/fcrepo4/tree/NewInjectedSessions/" https://www.pivotaltracker.com/story/show/4920564916:38
* ff-github joins
[fcrepo-fixity] fasseg pushed 1 new commit to master: http://git.io/SnbNFA
fcrepo-fixity/master 7b71ae0 fasseg: added test dependency for fcrepo-object-serialization
* ff-github leaves
<bljenkins>Yippie, build fixed!16:43
Project fcrepo-fixity build #272: FIXED in 4 min 43 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/272/
* github-ff joins16:44
[fcrepo4] ajs6f pushed 1 new commit to NewInjectedSessions: http://git.io/b63qmw
fcrepo4/NewInjectedSessions e86270f ajs6f: All endpoints save FedoraTransactions are injected, remains to add unit tests
* github-ff leaves
* nbanks joins16:49
<fasseg>hmmm every datastream I add to an object got info:fedora/fedora-system:def/internal#status BAD_CHECKSUM and BAD_SIZE..see http://fcrepo4.fcrepo.org/fcrepo/rest/objects/75a8dfda-0e58-40f5-b09d-9d4e176f4db9/license-1/fcr:fixity
<pivotal-bot>Andrew Woods finished "Update Cloud Tooling to support Provisioned IOPS" https://www.pivotaltracker.com/story/show/5156739116:51
<fasseg>can you guys verify that, or am I doing something wrong?
<pivotal-bot>Andrew Woods added comment: "Resolved with: https://github.com/futures/fcrepo-aws-puppet/commit/4155bd60163eb95a67ec49f813b197188161958c" https://www.pivotaltracker.com/story/show/5156739116:52
Andrew Woods delivered "Update Cloud Tooling to support Provisioned IOPS" https://www.pivotaltracker.com/story/show/51567391
<cbeer>fasseg: might want to ping barmintor.16:53
<fasseg>barmintor: Im getting BAD_CHECKSUM and BAD_SIZE for every datastream I add, e.g. a plain text file at: http://fcrepo4.fcrepo.org/fcrepo/rest/objects/75a8dfda-0e58-40f5-b09d-9d4e176f4db9/license-1/fcr:fixity Am I doing sth wrong or is this a bug?16:55
* github-ff joins
[fcrepo-aws-puppet] none pushed 3 new commits to master: http://git.io/TGsztw
fcrepo-aws-puppet/master 4155bd6 Andrew Woods: Update cloud-init scripts to allow for setting AWS Provisioned-IOPS and Optimized-EBS....
fcrepo-aws-puppet/master 9d064fa Andrew Woods: Rename README to README.md
fcrepo-aws-puppet/master 943a4b9 Andrew Woods: Update README.md
* github-ff leaves
<fasseg>cbeer: I guess by ping you mean mention his name directly or is he only reating to private messages?16:56
*reacting
<cbeer>fasseg: or via email. he's not on this sprint technically.
<fasseg>kk
* nbanks leaves16:57
<pivotal-bot>Andrew Woods started "Enable remote YourKit profiling with Puppet" https://www.pivotaltracker.com/story/show/5156790116:58
* bljenkins leaves
<ajs6f> Hey, git-meisters. I'd like to squash my commits before making a PR from this injected-sessions branch.16:59
But I've already pushed it non-local to github.
<cbeer>don't worry about it.
<ajs6f>Should I just rebase and squash locally, and then push it again?
<cbeer>yes
<ajs6f>OK
Thanks!
I guess it only matters if someone else has started off from my branch?
<cbeer>you'll break my local tracking branch.. but these feature branches should be ephemeral anyway.
<ajs6f>Right.
* github-ff joins17:04
[fcrepo4] ajs6f force-pushed NewInjectedSessions from e86270f to a1f4ed7: http://git.io/pBY7mQ
fcrepo4/NewInjectedSessions a1f4ed7 ajs6f: Field injection of JCR Sessions into JAX-RS endpoints, except FedoraTransactions
* github-ff leaves
* ff-github joins
[fcrepo-fixity] fasseg pushed 1 new commit to master: http://git.io/_9yFzw
fcrepo-fixity/master a890f1f fasseg: moved web.xml
* ff-github leaves
* bljenkins joins17:05
* github-ff joins17:06
[fcrepo4] ajs6f opened pull request #79: Field injection of JCR Sessions into JAX-RS endpoints, except FedoraTransactions (master...NewInjectedSessions) http://git.io/XL-NoQ
* github-ff leaves
<pivotal-bot>A. "Torbulater" Soroka added comment: "https://github.com/futures/fcrepo4/pull/79" https://www.pivotaltracker.com/story/show/4920564917:07
A. "Torbulater" Soroka finished "Inject authenticated sessions into the JAX-RS resources" https://www.pivotaltracker.com/story/show/49205649
<bljenkins>Project fcrepo-fixity build #273: SUCCESS in 3 min 47 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/273/17:08
<pivotal-bot>A. "Torbulater" Soroka added comment: "This work accomplishes Session injection for every endpoint in fcrepo4 except FedoraTransactions,..." https://www.pivotaltracker.com/story/show/4920564917:09
<cbeer>ajs6f: is there a good place for documentation about how and why that works17:10
?
<ajs6f>The injection thing?
_i_ barely know how it works. :)
Seriously...
<cbeer>yeah... or is it an obvious property of @Scope("prototype") classes?
(and if i had any clue what that meant, i'd understand what's happening)
<ajs6f>It's not so much a property of that scope, it's that injecting a request-scoped thing (like a Session) into a singleton bean makes no sense. So we had to make the scope of the endpoints shorter-lived.17:11
We could eventually do request-scoping or somtehing.
Just not anything that crosses request bounaries.
<aawoods>http://static.springsource.org/spring/docs/3.0.0.M3/reference/html/ch04s04.html
<ajs6f>thanks aawoods.17:12
Although scopes are a notion common to Spring and many other DI frameworks.
And full-bore JEE.
* jcoyne leaves17:13
<cbeer>ok.. and there's something magic in jersey or jax-rs or something that creates these new objects for each request?17:14
<ajs6f>Meaning the Sessions?17:16
)that get injected)
afk bbi317:17
back17:20
If you do mean the Sessions, it Jersey that does it (although it's spec in JAX-RS 2.0) and we impl their SPI.17:21
So when Jersey sees the @InjectedSession annotation, it knows
to ask for a InjectableSession from our provider.
The inejctable Session knows how to build a Session from HTTP context.
<cbeer>ah. so there's magic buried in that annotation17:22
<ajs6f>They do the hokey-pokey and they turn themselves around, and that's what it's all about.
It's DECLARATIVE!
Jersey is acting on the annotation, but it does that by executing our code.
We could inject anything. Annnnyythinggg.17:23
We could inject eddies into an endpoint.
<cbeer>ajs6f: i think you're safe to change TestHelper. I think I initially pushed that session logic in there after seeing it repeated N+1 times.
<ajs6f>Okay. I'll take your word for it.
But I'll do it tomorrow.
I'm heading out.
See y'all later.
<cbeer>k
* ajs6f leaves
<cbeer>aawoods: without objection, i'm going to merge that in now17:24
at least to avoid annoying merge conflicts later
looks like a lot of reformatting happened in that branch too
* jcoyne joins18:00
<pivotal-bot>Chris Beer delivered "Inject authenticated sessions into the JAX-RS resources" https://www.pivotaltracker.com/story/show/4920564918:04
* github-ff joins
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/zjtNLQ
fcrepo4/master d684999 Chris Beer: Merge pull request #79 from futures/NewInjectedSessions...
* github-ff leaves
<pivotal-bot>Chris Beer added "alter FedoraTransactions so that it no longer maintains the state of the transaction store inside the endpoint itself" https://www.pivotaltracker.com/story/show/51602513
Chris Beer estimated "alter FedoraTransactions so that it no longer maintains the state of the transaction store inside the endpoint its..." https://www.pivotaltracker.com/story/show/51602513
Chris Beer edited "alter FedoraTransactions so that it no longer maintains the state of the transaction store inside the endpoint itself" https://www.pivotaltracker.com/story/show/51602513
Chris Beer finished "Stop recreating a test namespace every time a JAX-RS resource is created." https://www.pivotaltracker.com/story/show/5157117318:10
* bljenkins leaves18:12
<pivotal-bot>Chris Beer estimated "Stop recreating a test namespace every time a JAX-RS resource is created." as 0 points https://www.pivotaltracker.com/story/show/5157117318:16
* github-ff joins
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/pfTwOw
fcrepo4/master 5f087c1 Chris Beer: move test namespace into our cnd
* github-ff leaves
* bljenkins joins18:17
<aawoods>cbeer: thanks for the heads up18:26
<cbeer>and then didn't wait for an answer :P18:27
<aawoods>I was away. That was reasonable.
* github-ff joins18:29
[fcrepo4] fasseg created tx-service (+1 new commit): http://git.io/Z0PXEA
fcrepo4/tx-service ed45fe1 fasseg: added TransactionService/-Tests in kernel and changed FedoraTransactions Endpoint to use the service
* github-ff leaves
<bljenkins>Yippie, build fixed!18:30
Project fcrepo4 build #803: FIXED in 26 min: http://ci.projectblacklight.org/jenkins/job/fcrepo4/803/
A. Soroka: Field injection of JCR Sessions into JAX-RS endpoints, except FedoraTransactions
Project fcrepo-fixity-corrupter build #89: SUCCESS in 43 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/89/18:31
<pivotal-bot>Frank Asseg added comment: "Created this branch: https://github.com/futures/fcrepo4/tree/tx-service
- Added TransactionService and a Un..." https://www.pivotaltracker.com/story/show/51602513
<bljenkins>Project fcrepo-fixity build #274: UNSTABLE in 2 min 38 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/274/18:33
* github-ff joins18:34
[fcrepo4] fasseg opened pull request #80: Pull request for https://www.pivotaltracker.com/story/show/51602513 (master...tx-service) http://git.io/mJrOAA
* github-ff leaves
<pivotal-bot>feature: alter FedoraTransactions so that it no longer maintains the state of the transaction store inside the endpoint itself (unstarted) / owner: Frank Asseg
* travis-ci joins
[travis-ci] futures/fcrepo4#683 (master - d684999 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/a35afeed07a5...d684999913b8
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/8032796
* travis-ci leaves
<pivotal-bot>Frank Asseg added comment: "Added pull request" https://www.pivotaltracker.com/story/show/5160251318:35
Frank Asseg added comment: "Added pull request: https://github.com/futures/fcrepo4/pull/80
" https://www.pivotaltracker.com/story/show/51602513
Frank Asseg finished "alter FedoraTransactions so that it no longer maintains the state of the transaction store inside the endpoint itself" https://www.pivotaltracker.com/story/show/51602513
* travis-ci joins18:36
[travis-ci] futures/fcrepo4#684 (master - 5f087c1 : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/d684999913b8...5f087c1067eb
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/8033172
* travis-ci leaves
* jcoyne leaves
<bljenkins>Project fcrepo-fixity-corrupter build #90: SUCCESS in 43 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/90/18:43
Project fcrepo-kitchen-sink build #375: STILL UNSTABLE in 2 min 42 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/375/18:46
Yippie, build fixed!
Project fcrepo-fixity build #275: FIXED in 3 min 48 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/275/
* github-ff joins18:53
[fcrepo4] fasseg force-pushed tx-service from ed45fe1 to 2aed1ac: http://git.io/X2kezA
fcrepo4/tx-service 2aed1ac fasseg: added TransactionService/-Tests in kernel and changed FedoraTransactions Endpoint to use the service
* github-ff leaves
* nbanks joins19:19
<bljenkins>Project fcrepo-fixity-corrupter build #91: SUCCESS in 1 min 24 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/91/
* nbanks leaves19:24
* fasseg leaves19:25
* github-ff joins
[fcrepo4] cbeer pushed 1 new commit to master: http://git.io/3fwLOg
fcrepo4/master 06b61ac Chris Beer: Merge pull request #80 from futures/tx-service...
* github-ff leaves
<bljenkins>Project fcrepo-kitchen-sink build #376: STILL UNSTABLE in 7 min 12 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/376/19:27
Project fcrepo-fixity build #276: SUCCESS in 10 min: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/276/19:28
<pivotal-bot>Edwin Shin started "Publish JMeter reports to Sonar via Maven" https://www.pivotaltracker.com/story/show/5151775319:33
Edwin Shin added comment: "jmeter-maven-plugin: http://jmeter.lazerycode.com/" https://www.pivotaltracker.com/story/show/5151775319:34
* travis-ci joins19:44
[travis-ci] futures/fcrepo4#686 (master - 06b61ac : Chris Beer): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/5f087c1067eb...06b61ac858e6
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/8034891
* travis-ci leaves
<bljenkins>Project fcrepo-fixity-corrupter build #92: SUCCESS in 35 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/92/19:46
Project fcrepo-fixity build #277: UNSTABLE in 2 min 48 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/277/19:48
Project fcrepo-kitchen-sink build #377: STILL UNSTABLE in 2 min 26 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-kitchen-sink/377/
* bljenkins leaves20:45
* bljenkins joins20:46
<aawoods>cbeer: Why did you take out the minutes from our sprint-11 close meeting? https://wiki.duraspace.org/display/FF/2013-06-11+Sprint+11+Close21:10
<cbeer>aawoods: the other sprints had a separate page about the retrospective21:11
so i moved them to a subpage there.
<aawoods>hmm
ok
<cbeer>i think that top page is for the report to the technical group21:12
<aawoods>so maybe we should summarize the outcomes of sprint-11
<cbeer>yeah, i was working on that21:13
<aawoods>I am not a big fan of just copying and pasting the entire body of other pages as the outcome, however.
It may be good to just list the exact outcomes of sprint-11...
not all of the outcomes up until and including sprint-11
<cbeer>aawoods: feel free to take a stab at it, then.21:15
<aawoods>cbeer: sure, tomorrow
* nbanks joins21:20
* nbanks leaves21:24
* github-ff joins21:47
[fcrepo4] eddies pushed 1 new commit to master: http://git.io/VGS96Q
fcrepo4/master d3ec209 Edwin Shin: 51517753: Publish JMeter reports to Sonar via Maven...
* github-ff leaves
<pivotal-bot>Edwin Shin added comment: "first pass: https://github.com/futures/fcrepo4/commit/d3ec2098bc6c5a97fe70d59c091bb6706499354d" https://www.pivotaltracker.com/story/show/5151775321:48
Edwin Shin added comment: "Added Jenkins job that uses the new jmeter profile: http://ci.fcrepo.org/jenkins/view/FF/job/fcrepo4-jmeter/" https://www.pivotaltracker.com/story/show/5151775321:49
<cbeer>cool
<aawoods>nice, eddies
<eddies>well, no idea if the reporting to sonar works yet21:50
but it's getting there
<aawoods>I was talking about you being awake ;)
<eddies>yeah. well. i'll pay for that later
<pivotal-bot>Edwin Shin added comment: "Sonar JMeter plugin 0.2 installed: http://sonar.fcrepo.org/updatecenter" https://www.pivotaltracker.com/story/show/5151775321:54
* travis-ci joins22:05
[travis-ci] futures/fcrepo4#687 (master - d3ec209 : Edwin Shin): The build passed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/06b61ac858e6...d3ec2098bc6c
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/8037724
* travis-ci leaves
* jcoyne joins22:08
<bljenkins>Project fcrepo-fixity build #278: FAILURE in 11 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity/278/22:22
Project fcrepo-fixity-corrupter build #93: FAILURE in 11 sec: http://ci.projectblacklight.org/jenkins/job/fcrepo-fixity-corrupter/93/
* bljenkins leaves
* bljenkins joins22:27
* bljenkins leaves22:37
* github-ff joins22:38
[fcrepo-aws-puppet] awoods pushed 1 new commit to master: http://git.io/I5bp0A
fcrepo-aws-puppet/master 92c74e2 Andrew Woods: Enable remote profiling with YourKit...
* github-ff leaves
* bljenkins joins
<pivotal-bot>Andrew Woods added comment: "Resolved with: https://github.com/futures/fcrepo-aws-puppet/commit/92c74e2e89b87fb3a51b3318b2b65bf2bfc9264d..." https://www.pivotaltracker.com/story/show/5156790122:41
Andrew Woods accepted "Enable remote YourKit profiling with Puppet" https://www.pivotaltracker.com/story/show/51567901
* jcoyne leaves22:58
<bljenkins>Project fcrepo4-jmeter build #1: SUCCESS in 19 min: http://ci.projectblacklight.org/jenkins/job/fcrepo4-jmeter/1/23:07
* nbanks joins23:20
* nbanks leaves23:25

Generated by Sualtam