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

Using timezone: Eastern Standard Time
* kaarefc joins04:00
* fasseg joins05:07
* fasseg leaves05:20
* fasseg joins05:34
* fasseg leaves06:28
* fasseg joins07:57
* jonathangee joins08:21
* fasseg leaves09:08
* gregjansen joins09:24
quick question for you guys.. In 3.x are there any implications for object state, beyond just metadata alone? (Not using CMA validation)09:26
We are debating using the "D" state to indicate a "move to trash"
* ksclarke joins09:28
* fasseg joins09:43
awoods: Im having trouble with the aws-puppet test.sh script you supplied. Can you take a look at this: https://gist.github.com/fasseg/7197100 ? the key error seems to be: "Could not evaluate: Could not retrieve information from environment production source(s) file:/tmp/fcrepo.war" Do I have to copy fcrepo.war by hand to /tmp or is this done by puppet?09:49
<awoods>hello, fasseg. I am taking a look.09:51
<fasseg>hey awoods... hope I dont have to upload it by hand, gonna take ages form my machine ;)
<awoods>no you do not09:52
<fasseg>only 500kb/s upstream :/
* osmandin joins09:54
<awoods>fasseg: It looks like this may be the issue: https://github.com/futures/fcrepo-aws-puppet/blob/master/site/site.pp#L909:55
fasseg: Probably need to update the URL to jenkins for the war download.09:56
<fasseg>I see
will do....
<pivotal-bot>Frank Asseg added comment: "yeah I had an error in there which I fixed on friday. So thorupghput will be reported accuratly." https://www.pivotaltracker.com/story/show/5925284009:57
<fasseg>err just a question about tickets, the requester "delivers" tickets, right?09:58
and awoods will accept...?
<awoods>fasseg: Yes, requester delivers, and owner finishes. "Deliver" means pushing the code/documentation to master or the wiki (which is usually the requester). I will "accept" at this point.10:00
<fasseg>k thx
<awoods>fasseg: The way it has been lately, is that the owner "finishes", and I have done the rest. But if we have another requester than myself, I would be happy for them to "deliver" when they are happy with the product, and I will "accept" after my review.10:01
<fasseg>just looking at the results of osman anyway....10:02
so I can deliver
If you are happy
<fasseg>it seems he has much better performance than me on the SCC cluster, which is very nice
<awoods>fasseg: that is a relief
<fasseg>~47mb/s with 4nodes/7threads10:03
could still be better....hope scott has some news about the cluster a wisconsin
<awoods>or gregjansen's cluster at UNC10:04
<pivotal-bot>Frank Asseg delivered "Create fcrepo4 single node benchmarking result on the Yale cluster" https://www.pivotaltracker.com/story/show/59252840
<awoods>or fasseg's cluster on AWS
<fasseg>if I ever get it running, I tried to day from the office., but ssh is even blocked in the "Guest" network, so I couldn't do **** there10:05
<gregjansen>meeting today with sysadmins for nodes in VM farm.. how many to ask for? I was going for 4
<fasseg>7 would be nice ;)
<awoods>fasseg: It is an odd situation: tasked with a job, then having your hands tied.10:06
<fasseg>but 4 is good too, I ran the SCC tests using 7 nodes, but osmandin used 4 nodes as well on the Yale cluster
<gregjansen>ok, 4 is probably more realistic here, but I will see
<fasseg>awoods: lesson learned: If I ever start my own company, delveloper will get their own subnet with no restrictions at all :)10:07
btw: I will need 2 days for SCAPE in the net sprint....
best would be 2 days sometime next week
* kaarefc leaves10:08
<awoods>fasseg: That could work out fine... as next week will be face-to-face hackery in Austin.
<fasseg>right I forgot...
<awoods>good luck, gregjansen, it your server acquisition.10:09
* github-ff joins10:13
[fcrepo-aws-puppet] fasseg pushed 1 new commit to master: http://git.io/4-XXTA
fcrepo-aws-puppet/master f7225a2 fasseg: updated WAR file location
* github-ff leaves
<fasseg>awoods: I updated the location of the war file and I can fetch it using wget 'http....' but I stil get this error message when running ./test.sh --noop: https://gist.github.com/fasseg/719746010:14
<awoods>fasseg: See comment in github10:16
fasseg: Otherwise, you may need to get your hands dirty in puppet.
* github-ff joins10:19
[fcrepo-aws-puppet] fasseg pushed 1 new commit to master: http://git.io/Vp67pQ
fcrepo-aws-puppet/master 8055b72 fasseg: added escape for '$' in war file path
* github-ff leaves
<fasseg>doesn't seem to help though.... I'll dig into it...10:20
<awoods>fasseg: are you using "sudo"?10:21
<fasseg>I did not on the last run :) but I does not make a difference....what about the hardcoded ip in fcrepo-aws-puppet/modules/tomcat7/lib/facter/private-ip.rb ...shouldn't this be lcoalhost or sth pointing to the puppet server? https://github.com/futures/fcrepo-aws-puppet/blob/master/modules/tomcat7/lib/facter/private-ip.rb10:23
<awoods>fasseg: No, that is an Amazon-global ip for getting info. Try it out on the instance.10:24
<fasseg>ah ok10:25
was bewildered by the 169.254 and thought this was class B or sth
* tecoripa joins10:28
<barmintor>awoods: Weekend was a little more hectic than I anticipated, but am running the release now.
<awoods>barmintor: late night at the disco?
<barmintor>awoods: lots of errands, and not a lot of energy :(10:29
<fasseg>awoods: can you paste your /etc/puppet.conf file?
* kaarefc joins
<fasseg>or rather /etc/puppet/puppet.conf10:30
gregjansen: oh and btw it would be nice to run at least two benchmarks on a cluster with a different number of nodes, so we can have a guess at the scalability factor...10:31
<gregjansen>fasseg: so at least two test runs on a cluster of a size not tested yet?10:32
<awoods>fasseg: I have no puppet.conf10:33
<fasseg>awoods: ah ok
<gregjansen> fasseg: so osman has tested on 4 and you have tested on 7, perhaps I should try sizes 3 and 5?
<tecoripa>fasseg, gregjansen: I want to talk to you two about clusters today... I think I can get us four VMs (or however many we decide) in in our library shop.
<awoods>fasseg: But I would definitely paste it if I did.
<fasseg>gregjansen: nah I meant just two runs on the same lcuster with diff sizes, you dont have to care about our cluster sizes
<gregjansen>fasseg: oh I see okay10:34
<tecoripa>fasseg: is there a wiki page describing the configuration of the clusters (number of VMs, RAM, CPU, disk space, network speed between nodes?)
<fasseg>and if possible it would be nice to have the same number of nodes as either osmandin or I had in the tests to also have a guess how the different cluster performances compare
<gregjansen>fasseg: so 4 and 710:35
<fasseg>tecoripa: you can take a look here, here are some infos: https://wiki.duraspace.org/display/FF/Performance+evaluation+on+the+SCC+Cluster
gregjansen: that'd be great yes
tecoripa: disk space was 500G/VM the rest should be on the page10:36
<tecoripa>fasseg: that's the SCC cluster, right?
yale cluster results from osmandin are here:
<tecoripa>fasseg: ok. and those are the only two clusters currently in use, right?
<fasseg>let me check for /proc/cpu on a SCC cluster VM
<awoods>tecoripa: yes, those are the only two clusters at the moment10:38
<fasseg>currently yes, besides the homemade cluster I have at home
but gregjansen is working on another one too...
<tecoripa>fasseg, awoods: ok. I can request some VMs here (easier and faster than going through the ACI, which is more for parallel computing), using the specs you put in the ACI request, fasseg. How many VMs would you like? Would you like fries with that?10:39
<awoods>tecoripa: I would say 7 VMs, so we have parity with SCC. fasseg?10:40
<fasseg>ill take as many as I can get ;) if you can get 20 ill take 'em....yes, with belgian sauce please and don't tell my GF ;)
* kaarefc leaves
* kaarefc joins10:41
<fasseg>well if we can have 20 we can always runa test with 7 active nodes
but I'd like to see how fedora4 handles 20, 50 or even 100 nodes
* ajs6f joins
<tecoripa>OK, I can ask for 7 VMs. Seriously, that may be an upper limit of what I can get quickly. Do we need 500GB per VM, for the tests?10:42
<fasseg>nah 100 is more than enough
yeah sure 7 is fine
<pivotal-bot>Andrew Woods added "Cluster-test top-level documentation" https://www.pivotaltracker.com/story/show/59668180
<awoods>tecoripa: Here is a ticket for creating a top-level cluster-testing description page, off of which SCC and Yale pages should hang. https://www.pivotaltracker.com/story/show/59668180
<pivotal-bot>feature: Cluster-test top-level documentation (unscheduled) / owner: Scott Prater
Andrew Woods edited "Cluster-test top-level documentation" https://www.pivotaltracker.com/story/show/5966818010:43
<tecoripa>fasseg, awoods: I can stub out a wiki page with the specs for the UW VMs, and use that as the base for my request. awoods, I'm thinking maybe a little wiki gardening is in order? Gather all the cluster info together, maybe organize it so it's easier to compare across clusters?
awoods: ah, you beat me to it.10:44
* ajs6f1 joins10:45
* ajs6f leaves10:46
<fasseg>oh US is not yet off daylight savings time so standup in 8 mins??10:52
* kaarefc leaves
<awoods>fasseg: correct
<awoods>fasseg: what does "kk" mean? Is it similar to "ok"?
<fasseg>awoods: yes it synonym, i think i picked that up while hanging out in IRC during my online gaming times (=university:)10:53
<awoods>fasseg: thanks. I may give it a spin.10:54
<fasseg>according to urban dictionary it stems form the gaming community
* tecoripa leaves10:55
<pivotal-bot>Andrew Woods edited "Parameterize JMS build port" https://www.pivotaltracker.com/story/show/5962831410:56
<pivotal-bot>A. "Senphitic" Soroka added "Alter HTML view machinery to deal with RDF iteration" https://www.pivotaltracker.com/story/show/59670062
<pivotal-bot>A. "Senphitic" Soroka added this label to 1 story https://www.pivotaltracker.com/story/show/58814930
A. "Senphitic" Soroka edited "Continue cleaning out unneeded static methods in JCR-Fedora contracts" https://www.pivotaltracker.com/story/show/58814930
<fasseg>btw are there plans to implement a SRU search provider?11:04
that'd be intersting for SCAPE since the APIs there use this spec...11:05
<ajs6f1>fasseg: I know of no concrete plans. Sounds like a good test of how easy (or not) it is to build an extension to the HTTP API.11:12
<fasseg>The extension works fine I did this already there: https://github.com/openplanets/scape-fcrepo4-connector and there https://github.com/openplanets/scape-fcrepo4-planmanagement11:13
<gregjansen>cbeer: your IDE bug might relate to path conflicts on Macs.. /tmp is a symlink on some macs, so canonical paths look different.
<cbeer>gregjansen: could be.. although this path is relative to ./target
<fasseg>ajs6f1: I could add the jax-rs resources just fine and inject the fcrepo4 services
<cbeer>i suspect my IDE uses a different CWD than maven11:14
once i finish this thread of work, i'll get the stacktrace
<ajs6f1>fasseg: Yeah. I'm not worried about _you_ doing it. I'm wondering if someone who isn't as knowledgable as you could do it without going crazy.
<gregjansen>perhaps the parent project's dir is your CWD?
cbeer: or vice versa
<ajs6f1>fasseg: I'm alo unenthused about the fact that you would have to package your module via Maven and not a proper modularization framework.11:15
<fasseg>ajs6f1: ah ok I misunderstood
hmmm that's a good point
<fasseg>I got so used to maven I didn't event think about that
<ajs6f1>I expect us to walk out of Austin with all of these things completely fixed, right?11:16
Well, fly out, anyway.
fasseg: normally, a deployer who wants to make the "SRU module" available shouldn't have to use dev machinery like Maven to do it, right?
<ajs6f1>I know my sysops would be unimpressed with that idea. {grin}
<awoods>cbeer: Using Intellij, you probably know that you can "Edit Configurations" of your test and set the "Working directory" to the module. By default the CWD is the top-level of the project.11:25
* tecoripa joins11:26
<cbeer>awoods: i probably don't know that.
<ajs6f1>barmintor/cbeer: GraphProperties… did we introduce that to account for user-supplied triple generation? IOW, would it be safe to factor it out of repo-default triple generation? I'm confident it will have no effect on testing, and right now it doesn't seem to do anything other than front for JcrRdfTools. I'm thinking I would factor it out of triple generation and break JcrRdfTools (which is huge) into two pieces, one for utility m11:27
<fasseg>awoods, tecoripa: I copied the fcrepo.war file by hand to /tmp and now im getting this error message from the test.sh script: 'Error: Failed to apply catalog: No title provided and "Notify[]" is not a valid resource reference'11:28
<awoods>cbeer: Checkout the drop-down of the test.
<fasseg>and im running this vs. a local puppet instance on my local box...right`
<tecoripa>fasseg: can you point me to the source you're using? the scripts, the puppet.conf?
<fasseg>tecoripa: sec..11:29
scripts are here: https://github.com/futures/fcrepo-aws-puppet in particular: https://github.com/futures/fcrepo-aws-puppet/blob/master/test.sh
<cbeer>ajs6f1: i think GraphProperties is there for the problems model stuff, used only on writes, i believe.
<ajs6f1>cbeer: That answers my question, because in the new modality, there is no "problems model" equivalent.11:30
<fasseg>my puppet.conf look likes this: https://gist.github.com/fasseg/7198902
<tecoripa>fasseg: I'll take a look
<ajs6f1>cbeer: So it should be fine for domain models to draw directly from JcrRdfTools (or its successor class).
<cbeer>ajs6f1: sounds good to me. i think the *RdfContext classes are probably the right level to inject user-supplied triple now11:31
(although we don't have a mechanism for that yet.. )
<fasseg>tecoripa: and the puppet log gives certificate errors: https://gist.github.com/fasseg/719894011:32
<ajs6f1>cbeer: No, but that would be the layer. A user could offer (via our awesome new modularization framework) a new RdfContext class which we would pick up and engage in the tree of iterators.
<fasseg>tecoripa: I guess I have to sign those certs first?
<ajs6f1>cbeer: k, I'm going to nuke GraphProperties for now.11:33
<cbeer>ajs6f1: sounds nice
<ajs6f1>cbeer: I figure you'll do the mod framework while cooking breakfast, right?
<tecoripa>fasseg: that looks familiar... let me look at the scripts more closely
fasseg: doesn't look like a signing issue: more like it cannot reach the master host11:34
<cbeer>ajs6f1: sure, no problem.
<fasseg>tecoripa: I thought I was running a master?
<cbeer>ajs6f1: i'm finding myself writing this logic a lot.. do you know if we have a helper method for this, or a suggestion where to put it?
ResourceFactory.createProperty(namespacedObject.getNamespaceURI(), namespacedObject.getLocalName()).asResource();
(converting from a JCR namespaced object into a Jena Resource)
<ajs6f1>"Would you like some bacon with that configuration?"
We have too many helper methods.11:35
(Too many to keep track of, that is.)
A namespaced object— would any Item be one of those?
<tecoripa>fasseg: puppet agent is the puppet client app. A host can be its own master and client.
<fasseg>tecoripa: I see...11:36
<ajs6f1>No, wait: you're talking about a Property -ish thing, right?
(JCR Property, I mean.)
<awoods>tecoripa/fasseg: This puppet set-up is stand-alone, no master node.
<cbeer>ajs6f1: nodes, properties, items, etc are all Namespaced
<fasseg>tecoripa: this command from the arch linux install guide seems to work: puppet facts find facter
<ajs6f1>cbeer: k
cbeer: No, I don't think we have a thing for that now. I would put it JcrRdfTools if I were you, and then I'll break it out of there into the former of the two classes I want to break JcrRdfTools into.11:37
<fasseg>awoods: stand-alone meaining it's a master, host and client all in one, right?
<ajs6f1>cbeer; That would be static, right?11:38
cbeer: Just thinking about tests.
<cbeer>ajs6f1: sounds good to me.
<tecoripa>fasseg: is puppet.localdomain and puppet in /etc/hosts?
<ajs6f1>cbeer: I do want to avoid setting us up to mock any more of our own statics.
to avoid that, I mean.
<fasseg>tecoripa: nah, want me to add them=
<tecoripa>yes, try that.11:39
also: what is $PUPPET_EXTRA_OPTS set to?
<fasseg>tecoripa: Not set11:40
<tecoripa>fasseg: can you do a ps, see if there is a puppet process running?
<fasseg>tecoripa: just the agent: root 25340 0.0 0.2 175372 34912 ? Ssl 16:39 0:00 /usr/bin/ruby /usr/bin/puppet agent11:41
adding the line to /etc/hosts didn't help
<awoods>fasseg: no, there is no master in this setup
<ajs6f1>cbeer: Maybe, instead of putting in JcrRdfTools, you can make a new class that holds the same context (session, GraphSubjects) as JcrRdfTools and has this translation and any others you end up needing as an instance method? That will make my life easier when you merge. ( I can break apart JcrRdfTools on my own time w/o worrying about conflicting with you.) And it avoids a mockable static.11:42
<tecoripa>fasseg: not sure how awoods set this up before, but I think you have to run puppet as a server: sudo puppet agent --server=puppet
(assuming "puppet" is now in /etc/hosts)
<fasseg>ill try that
<cbeer>ajs6f1: sure.11:43
<ajs6f1>cbeer: thnx!
<cbeer>bah. i thought i was doing such a good job writing my tests (and/or my code)11:45
<awoods>tecoripa/fasseg: You can run puppet in a serverless mode (no master), which is what is going on with this script. The AWS script allows you to start it locally, it spins up the AWS server and install/runs puppet on that instance without a puppet-master: http://docs.puppetlabs.com/learning/manifests.html#the-site-manifest-and-puppet-agent
<cbeer>but i forgot to assert anything :P
<ajs6f1>Invincible tests cannot fail!11:46
<fasseg>awoods, tecoripa: no luck, ps reports "/usr/bin/ruby /usr/bin/puppet agent --server=puppet" process running now, but I still get the same errors
<tecoripa>awoods, ah, okay. I'll look at that. I've always run it in a client/server mode.
<fasseg>awoods: I'll have to read up a bit on that I guess..
<awoods>tecoripa: See readme: https://github.com/futures/fcrepo-aws-puppet
<tecoripa>fasseg: revert everything to the way it was...
<fasseg>hehe k11:47
<tecoripa>awoods: yep, looking at that now
fasseg: let's see if we can get it running in the serverless mode.
<fasseg>awoods: So I dont need to have the local puppet agent service running? the scripts will spin one up?11:48
<awoods>fasseg: correct, no service required.11:49
fasseg: as long as puppet is installed, running "puppet apply" just asserts the scripts. period.11:50
<fasseg>awoods: k
ill try running the puppet command in test.sh ba hand then11:52
<tecoripa>fasseg: yeah, I think that's approach to go through right now: run the commands in create-ec2-instance.sh by hand, line by line, and make sure everything behaves as expected (and that all the dependencies are in place).11:55
<fasseg>tecoripa: Im not even do far as to run the actual ec2 instance creation script im still verifying the install using the test.sh script11:56
so this command fails: puppet apply --modulepath modules/ -v --noop modules/tomcat7/tests/init.pp
* osmandin leaves
<awoods>tecoripa/fasseg: Yes, the top-level ./test.sh command should be used for local (non-AWS) testing11:57
<fasseg>awoods: ah init.pp also has the path to the war file, that might be it11:58
<tecoripa>fasseg: add a "-d" to that command to get debug output
<cbeer>ajs6f1: huh. modeshape and/or JCR use the same "URI" to mean either a mixin or a property
<awoods>tecoripa/fasseg: The ./test.sh is helpful for running on an AWS instance that did not install as expected.
<cbeer>gonna make for some funny rdf
<tecoripa>fasseg: also "-v" for verbose
<ajs6f1>cbeer: Where?11:59
<fasseg>awoods: So this is not a script which should be run locally on my box, but on a EC2 instance?!12:00
I thought this would check if locally everything is there for using puppet to create ec2 instances
<awoods>fasseg: there are two scripts
<cbeer>ajs6f1: here's the rdfs output so far: https://gist.github.com/cbeer/fb13c8dec157db782d4712:01
<awoods>fasseg: one starts an instance and installs everything (/cloud-init/create-ec2-instance.sh)
<cbeer>e.g. http://www.modeshape.org/1.0repository
<tecoripa>fasseg: some paths hardcoded here:12:02
<awoods>fasseg: the other script (./test.sh) should be run either on the AWS (for testing) or on your local machine/VM that has the prerequisites already installed.
<fasseg>awoods: ok so it makes sense to get this running locally without a VM, since it detemines if all the prerequisites for the scripts are there...12:03
<ajs6f1>cbeer: ouch, the poorly-formed URIs… but yeah, it looks like they don't distinguish. Can we put some logic in to do that and file a ticket?
<cbeer>ajs6f1: yeah, we have logic around somewhere, i just need to remember where it is.12:04
ajs6f1: it's a JCR screwup.
<awoods>fasseg: You can test locally if you like, but I would prefer not to tweak my local machine and just do testing on an AWS instance.
<ajs6f1>cbeer: I hope they fix that.12:05
<cbeer>ajs6f1: i mentioned it to rhauch ages ago.. i can't remember if he said it was on the docket, or an intractable probelm.
<awoods>fasseg: BTW, puppet is not creating or starting the AWS instance. That is done by the "ec2run" script.
<tecoripa>fasseg: also checkout the paths here: https://github.com/futures/fcrepo-aws-puppet/blob/master/modules/tomcat7/manifests/init.pp
and the IP setting here: https://github.com/futures/fcrepo-aws-puppet/blob/master/modules/tomcat7/lib/facter/private-ip.rb12:06
<fasseg>awoods: ok
<cbeer>ajs6f1: but we definitely have logic around somewhere to fix it up for our purposes.
<fasseg>tecoripa: already updated these...
<ajs6f1>cbeer: How can that be intractable!?! Unless he means it's spec'd and they can't move until the spec is fixed...
<fasseg>and the private-ip is an amazon ec2 thing
<tecoripa>fasseg: as awoods is saying, this looks to be configured to run on the AWS instance.
<cbeer>ajs6f1: it's definitely spec12:07
<ajs6f1>cbeer: Oh. I hope theyhurry up with 2.112:08
<cbeer>i just can't remember if he said they could fix it in JCR any time
<ajs6f1>cbeer: I'd love to see them catch up with generics, try-with-close, all kinds of stuff. They could even decide to come up to Java8 and quasi-lambda and interfaces-with-defaults (aka mixins).
cbeer: It's really quite out of date.12:09
<cbeer>ajs6f1: did we ever come up with a better way to mock e.g. NodeTypeIterator?12:20
(iirc, modeshape's impl is private/protected)12:21
<ajs6f1>cbeer: For that kind of thing, what i do is write to our own (post-generics, typesafe) iterators in kernel.util.iterators
(Write client code, that is).
Then you can mock them, instead.12:22
(That's another thing I'd like to see in the spec. Typesafe iterators.)
I don't think we have one for NodeTypes, yet. Look at the ones that are there, tho'. It's trivial to write one and the tests you can copy from the examples almost verbatim.12:23
<cbeer>drat. i hoped we'd had some breakthrough there
<ajs6f1>cbeer: No, the "mock static and final crap from MODE and ISPN" thing is still an albatross around our neck. And I don't think it's going away anytime soon.12:26
cbeer: The best we can do, I think, is not make it worse by mocking our own stuff.12:27
<awoods>ajs6f1: "mocking our own stuff", meaning by not creating statics that need PowerMocking?12:28
<ajs6f1>awoods: Right.12:29
<ajs6f1>Sorry if that was uclear.
<cbeer>ajs6f1++ thanks12:34
<ajs6f1>cbeer: Sure. I wish there was something easier, but that was the best I could come up with. final classes--12:35
* gregjansen leaves
* osmandin joins12:39
<tecoripa>did a pull from fcrepo4/master now, and tried to rebuild...12:48
getting out of memory errros when running fcrepo-kernel tests
here's my MAVEN_OPTS:
MAVEN_OPTS="-Xmx512m -XX:MaxPermSize=512m"
anyone else having problems? I suppose Jenkins/Travis are doing fine12:49
* github-ff joins12:50
[fcrepo-aws-puppet] fasseg pushed 1 new commit to master: http://git.io/DSOhPA
fcrepo-aws-puppet/master 92e7945 fasseg: added arch linux installation section to readme
* github-ff leaves
<awoods>tecoripa: https://wiki.duraspace.org/display/FF/First+Steps#FirstSteps-BuildingFedora4
tecoripa: Give the following a shot: MAVEN_OPTS="-Xmx1024m -XX:MaxPermSize=1024m" mvn clean install12:51
<ajs6f1>Is anyone testing this thing on an Atari ST? 'Cause we have a _big_ community.
<tecoripa>awoods: yeah, I originally had it at 1024m, when I first got the error. I'll try again.12:52
* github-ff joins12:53
[fcrepo4] cbeer force-pushed cnd-to-rdfs from 94e4301 to 1e57edf: http://git.io/akTbNg
fcrepo4/cnd-to-rdfs 1e57edf Chris Beer: convert CND => RDFS
* github-ff leaves
<awoods>tecoripa: Running build now to double check all is well.
<tecoripa>awoods: still getting the same error. I have 8GB on my workstation, most of which is unused.12:54
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 2.185 sec - in org.fcrepo.kernel.spring.ModeShapeRepositoryFactoryBeanTest
Exception in thread "main" Exception: java.lang.OutOfMemoryError thrown from the UncaughtExceptionHandler in thread "main"
<awoods>tecoripa: My build now dumps as well.12:55
<awoods>cbeer: -Dlicense.skip
<tecoripa>awoods: I blame cbeer ;)
<fasseg>awoods: success! I got the script to create an EC2 instance!12:56
<cbeer>awoods: do you know checkstyles skip too, off the top of your head?
<cbeer>could have guessed that :P12:57
* github-ff joins
[fcrepo-aws-puppet] fasseg pushed 1 new commit to master: http://git.io/HqcEGQ
fcrepo-aws-puppet/master d4e906b fasseg: updated path to newest fcrepo4 version
* github-ff leaves
<ajs6f1>awoods: Do you know how to turn off checkstyles _warnings_?12:58
<awoods>tecoripa: The build was dumping last week due to memory used in tests, before ajs6f1 updated the top-level pom: https://github.com/futures/fcrepo4/commit/4fb55c1150124e3f549ac6d2173f563ba6636302#diff-600376dffeb79835ede4a0b285078036
tecoripa: It looks like we either need to dial up that configuration or clean up the PowerMocking in our tests.12:59
ajs6f1: I do not, off hand (if there even is a way)
<barmintor>ajs6f1: I think someone has aked that ? re: checkstyle in every sprint13:00
<ajs6f1>awoods: oh, well. I can (and have) adjusted it in-IDE, but it's annoying to wade through all that "line too long" crap.
barmintor: Shall I make you a ticket? {grin}
<tecoripa>awoods: I'll try editing the pom, increasing the MaxPermSize...
<barmintor>tecoripa: can you not pass the permSize in with MVN_OPTS ?13:01
<ajs6f1>tecoripa: if you just increase the heap using surefire argline, be sure to add in the inherited argilne, because jacoco uses that to pass in some config for itself. If you don't, you will stop generated test coverage reports. I pushed another one-line commit a day later to fix that in master.
tecoripa: so you can just rebase to get that13:02
barmintor: Not for tests!
barmintor: That's what was biting us in the hinder.
<barmintor>ajs6f1: lame.
<ajs6f1>barmintor: Surefire forks for tests (and failsafe uses surefire).
barmintor: it's actually really flexible, although I agree that the default is maybe wacked.13:03
<tecoripa>ajs6f1: ok, checking the inherited argline, too
<ajs6f1>We've never really tweaked our Maven test framework. When I started to do that, I saw massive speedup.
<barmintor>ajs6f1: do you have to do it in the pom, or can you do it in your settings.xml?13:04
<ajs6f1>Either. I did it in the POM, so that people didn't have to take extra action. But you can play it a couple of different ways.
barmintor: We should probably look at how things are nbeing forked, too, because we can parallelize a lot of the tests… as long as they're safe to do that. And if they aren't, that sometimes tells us things we'd want to know about race conditions, etc.13:05
<tecoripa>ajs6f1: inherited argLine? inherited from where?13:06
* github-ff joins
[fcrepo4] ajs6f pushed 1 new commit to RDFIterationForMutation2: http://git.io/ALBrjQ
fcrepo4/RDFIterationForMutation2 f679d4c ajs6f: Moving iteration all the way into domain models (e.g. FedoraResource)
* github-ff leaves
<ajs6f1>tecoripa: Other plugin settings (like jacoco)
<tecoripa>ajs6f1: I see that the jacoco plugin uses the failsafe.argLine, but that argLine is never defined anywhere (as far as I can tell)13:11
<ajs6f1>tecoripa: It's part of the context.
tecoripa: Take a look at master to see how to use it.
<tecoripa>ajs6f1: ah, okay
ajs6f1: I've checked master out from github, I did a pull, everything is up-to-date, and I still get the memory errors...13:19
<fasseg>awoods: running one aws instance only does not cost anything, right? so I only burn money when I spin up a cluster...?13:20
<ajs6f1>tecorpa: Where/when?
<awoods>fasseg: No, it all costs.
<fasseg>awoods: then Ill creat an account at: http://aws.amazon.com/free/ so I can run single node tests a no cost?13:21
<awoods>fasseg: That would be great.13:22
<tecoripa>unning org.fcrepo.kernel.spring.ModeShapeRepositoryFactoryBeanTest
INFO 12:21:41.292 (LogFactory) SLF4J implementation located in the classpath; it will be used for logging
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 2.853 sec - in org.fcrepo.kernel.spring.ModeShapeRepositoryFactoryBeanTest
Exception in thread "main" Exception: java.lang.OutOfMemoryError thrown from the UncaughtExceptionHandler in thread "main"
Results :
Tests run: 285, Failures: 0, Errors: 0, Skipped: 1
awoods: but those are only micro instances....13:23
<tecoripa>ajs6f1: looks maybe to be a jacoco problem?
<fasseg>have to check what that means
<awoods>fasseg: exactly
* jonathangee leaves
<awoods>fasseg: It means they are great for running itty bitty processes
<fasseg>613MB ram :/
<ajs6f1>tecoripa: and your poms are exactly as in master?
<awoods>fasseg: You have rights to use the fedora AWS account13:24
<fasseg>100mb storage, that's a bit on the slim side :)
<tecoripa>[ERROR] Failed to execute goal org.apache.maven.plugins:maven-surefire-plugin:2.16:test (default-test) on project fcrepo-kernel: Execution default-test of goal org.apache.maven.plugins:maven-surefire-plugin:2.16:test failed: The forked VM terminated without saying properly goodbye. VM crash or System.exit called ?
[ERROR] Command was/bin/sh -c cd fcrepo4/fcrepo-kernel && /usr/java/jdk1.7.0_13/jre/bin/java -javaagent:/home/sprater/.m2/repository/org/jacoco/org.jacoco.agent/ -jar /home/sprater/projects/fedora/fedora4/fcrepo4/fcrepo-kernel/target/surefire/surefirebooter6785830629440184578.jar /home/sprater/
<awoods>fasseg: but thanks for being money-conscious
<fasseg>awoods: sure, I now but I thought I can get the same at no cost for 750 hours
<tecoripa>ajs6f1: yes, except for the change I made to bump up the MaxPermSize
<ajs6f1>tecoripa: pull that out.
tecoripa: I want to start from a known place.13:25
<tecoripa>right. I pulled it out, did a git diff, excatly the same as what's on master
same error.13:26
<fasseg>hmm but that means I could set up an ec2 free instance as a http proxy to get around GEMA restrictions in germany, hrhr
<awoods>ajs6f1: does your fcrepo/master build at the moment?
<awoods>ajs6f1: not mine
<ajs6f1>I can't do much here, because I can't replicate.
awoods: All you. Own it!13:27
<awoods>ajs6f1: I am in the same boat as tecoripa.
<ajs6f1>awoods: Right. You can see the error. I can't,.
<awoods>ajs6f1: indeed
<ajs6f1>awoods: Why did it build previously?
awoods: Is CI reporting any problems?13:28
<tecoripa>ajs6f1, awoods: according to the error output, it looks like it's the forked jacoco process that's bombing
awoods: does that match what you see?
<awoods>tecoripa: yes13:29
ajs6f1: https://travis-ci.org/futures/fcrepo4/builds
ajs6f1: It looks like this commit pushed something over the line: https://github.com/futures/fcrepo4/commit/4bf5197ae12a7f34a43cf24f7f3ce82ebe614d2013:30
<tecoripa>awoods, ajs6f1: I see that travis errored out on the last build... at the same point, same error
<ajs6f1>That's what turned jacoco back on.
We can pull it out, but we lose unit test coverage reports.13:31
<bljenkins>Project fcrepo-fixity-corrupter build #318: SUCCESS in 1 min 42 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/318/
<ajs6f1>Might as well, I guess.
* jonathangee joins13:34
* jonathangee leaves13:36
* github-ff joins13:52
[fcrepo4] cbeer force-pushed cnd-to-rdfs from 1e57edf to 6882284: http://git.io/akTbNg
fcrepo4/cnd-to-rdfs 0c68dd3 Chris Beer: add a typesafe NodeTypeIterator to wrap the upstream JCR NodeTypeIterator
fcrepo4/cnd-to-rdfs 6882284 Chris Beer: Convert JCR node type definitions to RDFS, and expose it as a REST API endpoint
* github-ff leaves
<bljenkins>Project fcrepo-fixity-corrupter build #319: SUCCESS in 1 min 2 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/319/13:54
<awoods>ajs6f1/tecoripa: Indeed, pulling out the following update allows the build to complete: https://wiki.duraspace.org/display/FF/Fedora+4.0+Core+Capabilities13:55
rather: https://github.com/futures/fcrepo4/commit/4bf5197ae12a7f34a43cf24f7f3ce82ebe614d20
<pivotal-bot>Chris Beer added comment: "https://github.com/futures/fcrepo4/pull/101" https://www.pivotaltracker.com/story/show/5372244913:56
Chris Beer finished "serializing NodeTypes as RDFS" https://www.pivotaltracker.com/story/show/53722449
Chris Beer added comment: "@ajs6f I believe this is in master now, correct?" https://www.pivotaltracker.com/story/show/5380496113:57
<tecoripa>awoods/ajs6f1: as does running maven with the "-Djacoco.skip=true" flag set.
awoods/ajs6f1: so we know the problem is a jacoco one, as suspected...13:58
awoods/ajs6f1: I've been doing some reading and testing things out, but I don't have any solution yet.
<pivotal-bot>A. "Senphitic" Soroka added comment: "yes." https://www.pivotaltracker.com/story/show/5380496113:59
<awoods>tecoripa: Yank that line from the top-level pom.xml, if you do not find another solution. Then we will need to find another way to get our test-coverage back.
* gregjansen joins
<awoods>tecoripa/ajs6f1: Coveralls is an option: https://coveralls.io/repos14:04
<gregjansen>looking at build out for cluster testing now.. any test platform specs? OS preference, do we need apache/mod_jk?14:07
<awoods>gregjansen: I would suggest a *nix OS
<tecoripa>awoods: looks like it oriented towards ruby codebases? for pay? and hosted on their server (enterprise offering still under construction)? not that any of those things are bad per se, mind you...14:08
<awoods>gregjansen: re: apache, do we need to specify all of the software we want installed to the IT group?
tecoripa: https://coveralls.io/docs/java14:09
<gregjansen>no, I am going to have sudo
getting 8 nodes
<awoods>gregjansen: nice work
* gregjansen leaves14:14
<bljenkins>Project fcrepo-fixity-corrupter build #320: SUCCESS in 59 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/320/14:17
<pivotal-bot>Scott Prater added comment: "I completed the tests, and Greg fixed the problems they uncovered. All tests now pass." https://www.pivotaltracker.com/story/show/5916255614:20
Scott Prater finished "Create integration tests for Fedora objects with ACLs" https://www.pivotaltracker.com/story/show/5916255614:21
<bljenkins>Project fcrepo-kitchen-sink build #565: SUCCESS in 5 min 11 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/565/14:23
<pivotal-bot>Chris Beer accepted "turn mixin membership into rdf:type assertions" https://www.pivotaltracker.com/story/show/53804961
<cbeer>awoods: i'm happy to eat in, as long as it doesn't interfere with copious amounts of BBQ.14:25
i'll pull up the grocery lists from the last meetings. they're probably 80% useful for austin.14:26
* jonathangee joins14:28
<awoods>cbeer: Thanks14:33
<pivotal-bot>Andrew Woods estimated "turn mixin membership into rdf:type assertions" as 0 points https://www.pivotaltracker.com/story/show/5380496114:40
Osman Din added comment: "Documented in item d.214:47
https://wiki.duraspace.org/display/FF/Test+-+Cluster+Setup+-+Yale" https://www.pivotaltracker.com/story/show/59252250
Osman Din finished "Check how much the "write-behind" setting in infinispan impacts the ingest performance" https://www.pivotaltracker.com/story/show/5925225014:48
* gregjansen joins14:54
<pivotal-bot>Scott Prater added comment: "Created as a child page of Test Platform: https://wiki.duraspace.org/display/FF/Test+Clusters" https://www.pivotaltracker.com/story/show/5966818015:06
<ajs6f1>afk bbl
* ajs6f1 leaves
<pivotal-bot>Scott Prater finished "Cluster-test top-level documentation" https://www.pivotaltracker.com/story/show/59668180
* gregjansen leaves15:07
* gregjansen joins15:14
* github-ff joins15:15
[fcrepo4] ajs6f pushed 1 new commit to RDFIterationForMutation2: http://git.io/Na9bkw
fcrepo4/RDFIterationForMutation2 123aa78 ajs6f: Pushed RDF mutation all the way through HTTP API, for replacement of properties
* github-ff leaves
<pivotal-bot>Andrew Woods accepted "Cluster-test top-level documentation" https://www.pivotaltracker.com/story/show/5966818015:16
Andrew Woods added comment: "@scottprater, Is the code available somewhere for review?" https://www.pivotaltracker.com/story/show/59162556
Scott Prater added comment: "See https://wiki.duraspace.org/display/FF/Basic+Role-based+PEP and https://wiki.duraspace.org/display/FF/Ac..." https://www.pivotaltracker.com/story/show/5924701415:17
Scott Prater finished "Document basic roles setup for acceptance test" https://www.pivotaltracker.com/story/show/59247014
Scott Prater added comment: "Talked this over with our sysadmins this morning, expecting a response back within a day or so." https://www.pivotaltracker.com/story/show/5942778415:23
Scott Prater started "Submit request to use University of Wisconsin - Madison Library Technology Group a request for seven VM nodes" https://www.pivotaltracker.com/story/show/5942778415:24
Scott Prater added comment: "https://github.com/gregjan/fcrepo-authz/compare/scott-ittests" https://www.pivotaltracker.com/story/show/5916255615:26
* gregjansen leaves15:34
* fasseg leaves15:45
<pivotal-bot>Scott Prater added "Complete Test Cluster specs page for Yale" https://www.pivotaltracker.com/story/show/5969949415:47
Scott Prater added "Complete Test Cluster specs page for Karlsruhe SCC" https://www.pivotaltracker.com/story/show/5969963615:48
Scott Prater added "Complete Test Cluster specs page for UNC" https://www.pivotaltracker.com/story/show/5969975015:49
Scott Prater added "Complete Test Cluster specs page for UW Madison" https://www.pivotaltracker.com/story/show/5969986415:51
Scott Prater added "Complete Test Cluster specs page for AWS" https://www.pivotaltracker.com/story/show/5969993015:52
* pivotal-bot leaves16:16
* pivotal-bot joins
* ajs6f1 joins16:37
<cbeer>ajs6f1: ping?16:39
<ajs6f1>all: I'm moving on with this iterate-all-the-RDFs stuff under the assumption that we do not want to allow users via the APIs to mutate properties in our namespaces or JCR namespaces or add or remove mixIns in those namespaces. (They can do that if they want via direct manipulation of the repo, but not via our APIs.) Does that jibe with everyone else's understanding?
cbeer: pung.
(That was a dented table tennis ball.)
<cbeer>ajs6f1: i'm trying to do all the clever things with guava and have gotten confused.
<ajs6f1>cbeer: Don't worry, that's the point. As long as you can confuse other people too, you're doing good. Seriously, what's up?16:40
<cbeer>I'm trying to construct a filter that takes either a NodeDefinition or a PropertyDefinition, and made a Predicate<ItemDefinition>
(both ND and PD extend ID)
<ajs6f1>Seems right-ish.
What's blowing back at you?16:41
<cbeer>and then I did a filter(forArray(nodeType.getDeclaredPropertyDefinitions()), myPredicate))
To get back a Iterator<ItemDefinition>?
<cbeer>sorry.. i was trying to figure out how to get the syntax error out of the IDE16:42
basically, forArray is giving me an Iterator<PropertyDefinition>
<ajs6f1>Oh, foo. That's annnoying. What's the error?
You should be able to do this without errors or warnings or "unsafe' annotations.16:43
<cbeer>java: no suitable method found for filter(com.google.common.collect.UnmodifiableIterator<javax.jcr.nodetype.PropertyDefinition>,com.google.common.base.Predicate<javax.jcr.nodetype.ItemDefinition>)
method com.google.common.collect.Iterables.<T>filter(java.lang.Iterable<T>,com.google.common.base.Predicate<? super T>) is not applicable
(no instance(s) of type variable(s) T exist so that argument type com.google.common.collect.UnmodifiableIterator<javax.jcr.nodetype.PropertyDefinition> conforms to formal parameter type java.lang.Iterable<T>)
method com.google.common.collect.Iterables.<T>filter(java.lang.Iterable<?>,java.lang.Class<T>) is not applicable
(no instance(s) of type variable(s) T exist so that argument type com.google.common.collect.UnmodifiableIterator<javax.jcr.nodetype.PropertyDefinition> conforms to formal parameter type java.lang.Iterable<?>)
oh, looks like it gave me the wrong filter() impl16:44
ok, all is right again.
<ajs6f1>Ah, yes. The long-running battle between Iterator and Iterable?
ajs6f1: and, yes, that jibes with my understanding16:45
<ajs6f1>I did some things in the barnch I'm working on with RdfStream that may give you some ideas about how to avoid that stuff.
cbeer: thanks. If we don't have that policy, updating-by-replacement becomes a baroque nightmare of everflowering promises and poisonous betrayal.16:46
<cbeer>ajs6f1: yep. we might do some whitelisted stuff to let people mess with a selection of triples, i suppose16:48
* gregjansen joins16:49
awoods: I was going to fork fcrepo-authz into futures. Is that the approach you prefer?
<ajs6f1>cbeer: I abstracted it as policy, but in the most simplistic ways, with predicates and so forth.
cbeer: I can definitely see this cross-connecting into what gregjansen and tecoripa are up to.16:50
cbeer: Actually, that could be a serious story to sell. "Controi access or mutation on a _predicate-by-predicate, type-by-type_ basis."16:51
<gregjansen>awoods: there is also a "github transfer ownership" option..
<ajs6f1>cbeer: fcrepo3 never managed to package a type-based framework for authZ, although it was never too hard to ryo and people did.
frameowkr = authZ framework16:52
<gregjansen>cbeer: that's something we do at UNC, controlling who can write which predicates
* jonathangee leaves
<cbeer>ajs6f1: i wasn't even thinking authz. simply some things you can't change with update-replacement triples.
<pivotal-bot>Gregory Jansen started "Move Auth PEP project into Futures" https://www.pivotaltracker.com/story/show/5935958416:54
<ajs6f1>cbeer: Yeah, I was already in the next atmospheric layer. That's how I earnt my architecture astronaut's badge.16:55
gregjansen: We're setting up to be able to do this without too much pain, as long as we can do one thing: determine whether policy applies independently of what the policy may be or decide.
gregjansen: Otherwise, we have to take the hit of policy execution _all the time_.16:56
<gregjansen>ajs6f: i.e., the fewer policies are in scope, the better
<ajs6f1>gregjansen: True enough, and the extreme is, zero. I'd like to be able to ask (from code in the core (or extensions) that is client of the authZ code) "Should I even bother asking you? Are you aware of the existence of this kind of decision?"16:58
gregjansen: Which ideally would be answered _very_ quickly.
<gregjansen>ajs6f1: this is one use of hierarchy I would think.. segment repository into domains of policies?16:59
limit "global policies"17:00
<ajs6f1>all: The question I asked earlier (to which cbeer gave the answer with which I'm going) has a child: are properties in _all_ of our namespace off-limits (at least by default, at least now)? I.e., there's:
(Probably shouldn't be off-limits.)
And then there's http://www.jcp.org/jcr/1.017:01
<tecoripa>I would think the first one should be off-limits to everyone but fedoraAdmin... but fedoraAdmin should be able to get to them.
<ajs6f1>gregjansen: That sounds very pragmatic.
gregjansen: IOW, ideally, I'd like to avoid loading too many semantics on the hierarchy, but it's the fastest horse we can ride.17:02
tecoripa: That's a "no". Because at this stage, my code doesn't know how to distinguish between Jane Doe and Fedo Radmin.17:03
(Or is that Raadmin? I can never remember, because I can't remember what strange country Fedo hails from.)
tecoripa: This is just the kind of thing that ought to be recorded in declarative policy, not imperative code, right?17:04
all: On a related note: our mixins are currently in http://fedora.info/definitions/v4/rest-api#17:05
Do we want them in http://fedora.info/definitions/v4/repository#
<tecoripa>just a sec...
<ajs6f1>It seems to me we do, but I'm not sure why they were in an API-related NS to begin with, so I tread carefully.
Also, I'm treading carefully because the rooster has come up here on the deck and he's looking angry. But then, he always looks angry.17:06
* jonathangee joins17:13
<cbeer>ajs6f1: jcp.org should be translated to v4/repository whenever it goes outside JcrRdfTools17:17
ajs6f1: beats my why models ended up in rest-api#, i saw that too and it seems wrong.
and rels-ext is fair game.
<ajs6f1>cbeer: Right.17:18
<cbeer>and predicates in repository# are almost certainly off-limits
<ajs6f1>cbeer: Except people could bring jcp.org in without having generating the relevant data from the repo, so it still matter.
IOW you could just walk up to a repo and barefaced, try to persist something with jcp.org in it.17:19
<cbeer>they should either be structural (so you can't change them with just a SPARQL-UPDATE) or derive from the running state (so can't be changed at all)
i think there's logic already in there to stop you from introducing a jcp.org/jcr/1.0 predicate
* cbeer i may bbe wrong
<barmintor>there was that logic17:20
but it's been so long...
<cbeer>yeah. ajs6f: see isInternalProperty
the LDP predicates also get special treatment
<ajs6f1>There was. I had to replace that, because it came from the Old World of "We have all the RDF in a query-able bag" and I'm exploring the New World of "You're throwing triples at me and I'm trying to juggle them without holding onto them for more than a moment."17:21
cbeer: Yep, got that handled.
(LDP, that is.)
(That fell into the same category as hierarchy-derived triples. It was obvious even to me that we aren't offering them for mutation.)
cbeer: where is isInternalProperty? I got a bad feeling I may have re-impl'd it.17:22
<ajs6f1>cbeer: Thanks. That's cool. I knew about that 'un.17:23
<barmintor>ajs6f1: I know you guys might've disabled it, but there is a copy-code report in … that thing with the test coverage
<ajs6f1>I did something in RdfLexicon (which should probably be moved away from there) which is a little more flexible.
* jonathangee_ joins
<ajs6f1>barmintor: We killed the test coverage, remember?17:24
* barmintor frowns
<ajs6f1>I agree.
That's why I reanabled it.
<ajs6f1>Sonar doesn't see branches, which bugs me. (Or so awoods told me.)
* jonathangee leaves17:25
<barmintor>right, but it has a copycode report
<ajs6f1>But I'm on a branch.
<barmintor>well, don't c&p code
<ajs6f1>barmintor: I DIDN'T.
<barmintor>didn't you?
<ajs6f1>barmintor: No. I read it, understood it, then re-impl'd the functionality more expansively.17:26
Not to be snarky or anything.
But I wasn't sure whether my code represented the policy with fidelity. Hence the preceding discussion.17:27
Actually, I wasn't sure that we had a policy. Now I am. And I'm sure that I did the right thing. Which I wasn't sure as I did it.17:28
* gregjansen leaves17:29
<ajs6f1>Now I'm going to expand my code to better represent the policy and offer a hook for future connection to declarative policy.
* jonathangee leaves17:32
<ajs6f1>On a sidebar, is there some agreement that our mixins belong in http://fedora.info/definitions/v4/repository# and not in http://fedora.info/definitions/v4/rest-api#?17:34
* jonathangee joins17:35
* jonathangee_ joins17:37
* jonathangee leaves17:40
* jonathangee_ leaves17:42
<pivotal-bot>Scott Prater started "Fedora performance tests with/without active basic roles and PEP enforcement" https://www.pivotaltracker.com/story/show/5941731417:43
* github-ff joins18:02
[fcrepo4] cbeer force-pushed cnd-to-rdfs from 6882284 to f87fd7b: http://git.io/akTbNg
fcrepo4/cnd-to-rdfs f87fd7b Chris Beer: Convert JCR node type definitions to RDFS, and expose it as a REST API endpoint
* github-ff leaves
* github-ff joins18:14
[fcrepo4] cbeer force-pushed cnd-to-rdfs from f87fd7b to 42204ac: http://git.io/akTbNg
fcrepo4/cnd-to-rdfs 42204ac Chris Beer: Convert JCR node type definitions to RDFS, and expose it as a REST API endpoint
* github-ff leaves
<cbeer>ajs6f1: i'm not sure how the zipping iterator fits into that work. feel free to fix it up if you want18:15
* ksclarke leaves18:19
<bljenkins>Project fcrepo-fixity-corrupter build #321: SUCCESS in 1 min 28 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/321/18:21
<ajs6f1>cbeer: I will take a look tomorrow. Time for my supper now. I'll be astonished if I can actually add any real quality to your branch, but from my point of view, life _is_ an endless joyful Hail-Mary-pass attempt to astonish ourselves, so perhaps I'll succeed.18:22
* ajs6f1 leaves
* tecoripa leaves18:26
* jonathangee joins18:31
* osmandin leaves18:32
* jonathangee leaves18:35
* jonathangee joins18:36
* jonathangee leaves18:39
<awoods>Seems like I missed some good activity in the last hour or two. barmintor: Looks like a release has breached.18:42
<bljenkins>Project fcrepo-fixity-corrupter build #322: SUCCESS in 1 min 14 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/322/18:44
Project fcrepo-kitchen-sink build #566: UNSTABLE in 3 min 43 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/566/18:47
<pivotal-bot>Andrew Woods accepted "Create fcrepo4 single node benchmarking result on the Yale cluster" https://www.pivotaltracker.com/story/show/5925284019:36
Andrew Woods added comment: "There is a significant difference between the Yale and " https://www.pivotaltracker.com/story/show/5925256419:37
Andrew Woods accepted "Create fcrepo3 benchmarking result on the Yale cluster" https://www.pivotaltracker.com/story/show/5925256419:38
Andrew Woods accepted "Check how much the "write-behind" setting in infinispan impacts the ingest performance" https://www.pivotaltracker.com/story/show/5925225019:39
* github-ff joins19:44
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/7nCZTw
fcrepo4/master bd7daf5 Andrew Woods: Re-disable Jacoco args, as they blow the build's permgen
* github-ff leaves
* travis-ci joins19:55
[travis-ci] futures/fcrepo4#1070 (master - bd7daf5 : Andrew Woods): The build was fixed.
[travis-ci] Change view : https://github.com/futures/fcrepo4/compare/4bf5197ae12a...bd7daf5f0ac2
[travis-ci] Build details : http://travis-ci.org/futures/fcrepo4/builds/13183934
* travis-ci leaves
* github-ff joins20:01
[fcrepo4] ajs6f pushed 1 new commit to RDFIterationForMutation2: http://git.io/5AD74g
fcrepo4/RDFIterationForMutation2 d07d83b ajs6f: Minor refactoring
* github-ff leaves
<bljenkins>Project fcrepo-fixity-corrupter build #323: SUCCESS in 1 min 35 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/323/20:03
Yippie, build fixed!20:12
Project fcrepo-jms-indexer-pluggable build #103: FIXED in 10 min: http://ci.fcrepo.org/jenkins/job/fcrepo-jms-indexer-pluggable/103/
<pivotal-bot>Andrew Woods accepted "Document basic roles setup for acceptance test" https://www.pivotaltracker.com/story/show/5924701420:18
Andrew Woods added comment: "Cloning: https://github.com/futures/fcrepo-authz.git20:32
Running: mvn clean install
Seeing: bui..." https://www.pivotaltracker.com/story/show/59162556
Andrew Woods rejected "Create integration tests for Fedora objects with ACLs" https://www.pivotaltracker.com/story/show/59162556
<bljenkins>Project fcrepo-fixity-corrupter build #324: SUCCESS in 1 min 52 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-fixity-corrupter/324/20:33
<pivotal-bot>Andrew Woods added comment: "The project: https://github.com/futures/fcrepo-authz, has three modules: ""20:37
a- fcrepo-auth-roles-basic
b- fcre..." https://www.pivotaltracker.com/story/show/59359584
<bljenkins>Yippie, build fixed!20:42
Project fcrepo-kitchen-sink build #567: FIXED in 9 min 17 sec: http://ci.fcrepo.org/jenkins/job/fcrepo-kitchen-sink/567/
* ksclarke joins22:07
* ksclarke leaves23:56

Generated by Sualtam