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

Using timezone: Eastern Standard Time
* dwilcox joins03:07
* dwilcox leaves03:34
* dwilcox joins03:40
* dwilcox leaves05:44
* ajs6f joins08:10
* awead joins08:20
* dhlamb joins08:41
* jcoyne joins08:43
* acoburn joins08:56
* escowles joins09:19
* MohamedAR joins09:32
* dwilcox joins09:33
<awoods>ajs6f: I think there are two PRs that are awaiting your final review:09:49
https://github.com/fcrepo4/fcrepo4/pull/782 and
* github-ff joins09:50
[fcrepo4] escowles pushed 1 new commit to default-identifier-translator: http://git.io/vf2DR
fcrepo4/default-identifier-translator 5e74e60 Esmé Cowles: Updating Javadoc comments
* github-ff leaves
* github-ff joins09:55
[fcrepo-audit] awoods closed pull request #4: Full URL form of PREMIS should be used for rdf:type (master...internal-audit) http://git.io/vfEQg
* github-ff leaves
* MohamedAR1 joins10:01
* MohamedAR leaves10:04
* travis-ci joins10:10
fcrepo4/fcrepo4#3619 (default-identifier-translator - 5e74e60 : Esmé Cowles): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/d9cc048865f3...5e74e600a4a6
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/59881211
* travis-ci leaves
* ksclarke leaves10:11
<f4jenkins>Project fcrepo-message-consumer build #1147: UNSTABLE in 4 min 47 sec: http://jenkins.fcrepo.org/job/fcrepo-message-consumer/1147/10:12
* whikloj joins10:31
* github-ff joins10:56
[fcrepo4] escowles pushed 1 new commit to default-identifier-translator: http://git.io/vfaOi
fcrepo4/default-identifier-translator de750f0 Esmé Cowles: Creating new PrefixingIdentifierTranslator and making DefaultIdentifierTranslator extend from that
* github-ff leaves
* ajs6f1 joins11:09
* ajs6f leaves
<awoods>ajs6f1/escowles: It would appear that the following ticket is ready to ship, no? https://jira.duraspace.org/browse/FCREPO-148911:11
<ajs6f1>I have no objection. I would like to see that TODO added to the PR, because we'll forget it otherwise.11:12
<escowles>awoods: almost: after fcrepo4 #781 and #782 are merged, i'll need to update fcrepo-audit #3 to use the new PrefixingIdentifierConverter
<ajs6f1>escowles: Thanks for doing that.
<escowles>ajs6f1: no problem!11:13
<awoods>escowles: I will squash/merge 781 now11:14
<escowles>awoods: excellent!
<f4jenkins>Yippee, build fixed!11:19
Project fcrepo-message-consumer build #1148: FIXED in 5 min 34 sec: http://jenkins.fcrepo.org/job/fcrepo-message-consumer/1148/
* travis-ci joins11:20
fcrepo4/fcrepo4#3622 (default-identifier-translator - de750f0 : Esmé Cowles): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/5e74e600a4a6...de750f0d55a2
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/59891561
* travis-ci leaves
* github-ff joins11:21
[fcrepo4] escowles pushed 1 new commit to default-identifier-translator: http://git.io/vfaBe
fcrepo4/default-identifier-translator 384d722 Esmé Cowles: Adding TODO note for ajs6f
* github-ff leaves
<escowles>ajs6f1: i added a note for you: https://github.com/fcrepo4/fcrepo4/commit/384d72279b265003454f73828a67f815b68cd01d
<ajs6f1>escowles: much obliged.
<f4jenkins>Yippee, build fixed!
Project fcrepo-camel-toolbox build #49: FIXED in 3 min 54 sec: http://jenkins.fcrepo.org/job/fcrepo-camel-toolbox/49/
* github-ff joins11:23
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/vfaRk
fcrepo4/master c5e15dd Esmé Cowles: Add FedoraResource.setURIProperty method...
* github-ff leaves
* github-ff joins11:24
[fcrepo4] awoods closed pull request #781: Adding FedoraResource.setURIProperty method (master...setURIProperty) http://git.io/vfBkf
* github-ff leaves
<awoods>escowles: 782 is now ready?11:25
<escowles>awoods: ship it!
* github-ff joins11:33
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/vfagY
fcrepo4/master 2c3eb0f Esmé Cowles: Update DefaultIdentifierTranslator to make it usable by audit service...
* github-ff leaves
* github-ff joins11:34
[fcrepo4] awoods closed pull request #782: Updating DefaultIdentifierTranslator to make it usable by audit service (master...default-identifier-translator) http://git.io/vfuwQ
* github-ff leaves
* github-ff joins11:35
[fcrepo-audit] escowles pushed 1 new commit to removeJCRleakage: http://git.io/vfa2k
fcrepo-audit/removeJCRleakage 7d9f771 Esmé Cowles: Using new PrefixingIdentifierTranslator class
* github-ff leaves
<escowles>awoods: i've updated https://github.com/fcrepo4-labs/fcrepo-audit/pull/3
* travis-ci joins
fcrepo4/fcrepo4#3624 (default-identifier-translator - 384d722 : Esmé Cowles): The build has errored.
Change view : https://github.com/fcrepo4/fcrepo4/compare/de750f0d55a2...384d72279b26
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/59895899
* travis-ci leaves
<awoods>escowles: #3 is ready?11:36
<escowles>awoods: yep
* MohamedAR joins11:37
* MohamedAR1 leaves
<escowles>afk # lunch11:40
<awoods>ajs6f1: what do you think about this: https://github.com/fcrepo4-labs/fcrepo-audit/pull/3/files#diff-3ff409abe3384fddfe4bc7d36c6ce055R26111:41
ajs6f1: I would rather fcrepo-audit not depend on fcrepo-kernel-impl11:42
ajs6f1: and that the identifierTranslator be injected...
<ajs6f1>awoods: I don't understand what that gist is supposed to show.
<awoods>ajs6f1: it is not a gist, but a line in escowles latest PR11:43
<ajs6f1>right, sorry— I still don't know what it means
auditResource.replaceProperties(new PrefixingIdentifierTranslator(session, baseURL + "/"), m,
+ new RdfStream(), containerService);
<awoods>ajs6f1: it is supposed to show the creation of the PrefixingIdentifierTranslator
<ajs6f1>Isn't that pretty much what escowles was already doing?
<awoods>ajs6f1: That is what he is proposing doing in this PR11:44
ajs6f1: my question relates to a direct dependency on fcrepo-kernel-impl
* jgpawletko joins
<ajs6f1>Yeah, it would be better to inject the identifier translator, but from where?
<awoods>ajs6f1: the translator would need a no-arg constructor11:46
ajs6f1: then setters for the session and namespace.
<ajs6f1>awoods: No, the translator can vary from request to request.
awoods: It's not a manageable bean.
<awoods>ajs6f1: yes, the setters would have to be called on each invocation
<ajs6f1>awoods: There's no point to that. That's not any better than what's ther now.11:47
<awoods>ajs6f1: you are probably right... shipping it.
* jgpawletko leaves
* travis-ci joins
fcrepo4/fcrepo4#3626 (master - c5e15dd : Esmé Cowles): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/fcec1046e2cd...c5e15ddf2bb5
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/59896281
* travis-ci leaves
<ajs6f1>awoods: I think you're right to be suspicous— I'm saying that there's a bigger issue behind this which is about how we manage identifier translators generally.
* github-ff joins11:48
[fcrepo-audit] awoods pushed 1 new commit to master: http://git.io/vfaK5
fcrepo-audit/master ace483a Esmé Cowles: Remove JCR leakage...
* github-ff leaves
* github-ff joins11:49
[fcrepo-audit] awoods closed pull request #3: Removing JCR leakage (master...removeJCRleakage) http://git.io/vfa6I
* github-ff leaves
* jgpawletko joins11:51
<f4jenkins>Project fcrepo4-release-tests build #200: UNSTABLE in 1 min 25 sec: http://jenkins.fcrepo.org/job/fcrepo4-release-tests/200/11:59
* pgwillia joins12:00
* travis-ci joins12:01
fcrepo4/fcrepo4#3627 (master - 2c3eb0f : Esmé Cowles): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/c5e15ddf2bb5...2c3eb0ffc615
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/59897779
* travis-ci leaves
<pgwillia>Does anyone have any advice about improving the performance of writing to fedora?
<ajs6f1>Faster hard drives.
<acoburn>local SSDs are the best (though harder to increase in size), if you're using a shared SAN, there can be latency issues12:02
<ajs6f1>Use LevelDB.12:03
<acoburn>pgwillia: is this fedora3 or fedora4?
<jcoyne> pgwillia: are you writing via active-fedora?
<pgwillia>jcoyne: yes12:04
<acoburn>binary data or metadata?
<jcoyne>pgwillia: Have you isolated the Solr writes?
<pgwillia>jcoyne: what do you mean by that?
<ajs6f1>This doesn't sound like it has anything to do with Fedora.12:05
<jcoyne>active-fedora is writing to Fedora & Solr.
<acoburn>does active fedora commit on each write to solr?
<jcoyne>so your performance benchmark isn't just measuring Fedora write speed.
acoburn: it does a softCommit
<acoburn>but it waits for the commit response with each write to solr? That will slow things down12:06
you could use a commitWithin value
which gives you better performance12:07
<ajs6f1>Does this belong in #projecthydra?
<acoburn>ajs6f1: yes it does
<ajs6f1>I don't mean to kick anyone out, but people in that room might benefit from hearing this.
<pgwillia>my question was specificaly about any advice about tuning fedora. The other stuff is useful.
<ajs6f1>pswillia: You need to isolate Fedora first before you know if you have any real need to tune Fedora.12:08
<pgwillia>ajs6fl: will do. I was hoping there might be some documentation I could be pointed at in the meantime.12:09
<ajs6f1>That I don't know.
<escowles>pgwillia: we do have some docs on the wiki about our performance testing: https://wiki.duraspace.org/display/FEDORA4x/Performance12:15
<escowles>the biggest impact changes are making sure your JVM has a large heap, and making sure you don't have a large number of child nodes under a single node
* acoburn leaves12:17
* github-ff joins12:28
[migration-utils] nianma opened pull request #7: https://jira.duraspace.org/browse/FCREPO-1413: apply and pass the checks... (master...fcrepo-1413) http://git.io/vfapL
* github-ff leaves
* acoburn joins12:34
* ksclarke1 joins12:51
* ksclarke leaves12:52
* ksclarke joins
<awoods>escowles: as you may be aware, any next release is currently on hold until these two b-node blockers are resolved:12:53
<escowles>awoods: fyi, i'm not getting events syncing to the external triplestore with fcrepo-camel-toolbox/audit-triplestore12:54
i'm getting java.lang.NoClassDefFoundError: Could not initialize class org.fcrepo.audit.InternalAuditor
both when i deploy the webapp or when i do mvn jetty:run12:55
<awoods>escowles: are you using the master branch of the toolbox?
<escowles>awoods: yes, are there outstanding PRs?
<awoods>escowles: none that are related.12:56
<escowles>hmmm... i wonder if acoburn or MohamedAR know why this isn't working?
<acoburn>escowles: I'm taking a look now12:57
<awoods>escowles: you may want to remove your local fcrepo-audit artifact?
<escowles>awoods: i'll try that
<acoburn>escowles: it will take me a few minutes, as I'm not at my usual computer (need to install fcrepo, fuseki, etc)12:59
<awoods>acoburn: or tweak vagrant?
<acoburn>that's a good idea13:00
<escowles>awoods: acoburn: flushing dependencies and rebuilding fcrepo-camel-toolbox didn't change anything13:03
<awoods>ruebot: is there an Islandora/F4 meeting now?
<escowles>the audit-triplestore-web WAR file contains the fcrepo-audit jar file (WEB-INF/lib/fcrepo-audit-4.1.2-20150424.155112-8.jar)13:04
<acoburn>... still downloading ...13:05
<escowles>acoburn: i stepped back to https://github.com/fcrepo4-labs/fcrepo-camel-toolbox/commit/5ec6b93c8bc480ad40a7b72b3709f32a7e53f6a3 and that works (both mvn camel:run and the war file)13:11
so it's definitely the refactoring to use InternalAudit for namespaces and getAuditEventType that changed things13:12
<awoods>escowles: we should have an integration test ;)13:14
<escowles>awoods: i know! that was next on my todo list if I had time...
<MohamedAR>escowles: Just saw the messages. I will try to find what's causing the problem.13:25
<escowles>MohamedAR: thanks -- i'm guessing something about the dependencies in your last commit13:26
<ruebot>awoods: yeah. i've got you on the list.13:31
awoods: sorry, for the delay. danny is talking now, and i have time to look around :-)13:32
<awoods>ruebot: I thought it started at 1?
<ruebot>awoods: yeah. it did. i had you on the skype call list.
<awoods>ruebot: I need to step away in 7 min13:33
<ruebot>awoods: you're yellow in skype. just tried adding you in.
<awoods>ruebot: so it is probably too late
* awoods wishing I could have just called in
<ruebot>awoods: ah, no worries. we've spent most of the time on this https://github.com/Islandora-Labs/islandora/blob/7.x-2.x/docs/technical-documentation/services.md
* ajs6f1 leaves13:38
<awoods>escowles: I need to step away until 3:30 or 413:44
<escowles>awoods: ok, i'm going to be out from 3:15 until probably 4, fyi13:45
* acoburn leaves14:03
* acoburn joins
* dhlamb leaves14:04
* dwilcox leaves14:06
* dwilcox joins14:07
* dhlamb joins14:11
<ruebot>awoods: here are the notes: https://github.com/Islandora/Islandora-Fedora4-Interest-Group/blob/master/meetings/06.md14:19
awoods: i'll give freeconferencecallHD a shot next time.14:20
* dhlamb leaves16:01
<awoods>escowles/MohamedAR/acoburn: please let me know if you have had success with the internal-audit flow.
<acoburn>awoods: on a call16:02
<awoods>acoburn: be safe16:03
<acoburn>awoods: I got as far as confirming the error exists and adjusting a unit test in the camel code to trigger the error16:12
<awoods>acoburn: ok... I am back and looking myself.16:13
acoburn: unit test or integration test?16:14
<acoburn>awoods: unit test in fcrepo-camel-toolbox/audit-triplestore
<acoburn>I think it had to do with the method returning null
<awoods>acoburn: Which class do you think has the issue? AuditSparqlProcessor?16:17
<acoburn>awoods: I take that back
I get: java.lang.NoClassDefFoundError: Could not initialize class org.fcrepo.audit.InternalAuditor16:18
<awoods>acoburn: where do you see that message?
<acoburn>mvn camel:run and then add an object to fedora
<awoods>acoburn: which application is throwing that error?16:19
acoburn: fcrepo? toolbox?
<acoburn>the error happens in fcrepo-camel-toolbox
because it can't load the class from fcrepo-audit16:20
<awoods>acoburn: maybe fcrepo-audit is not a proper OSGi bundle?
<acoburn>possibly, I started pursuing that path, but nothing stood out
actually, that may be the issue16:23
this is part of why none of the fcrepo-camel code depends on any of the fcrepo core modules
there's an issue with the javax.inject package, which is required by one of the dependencies of fcrepo-audit16:24
I need to leave for the weekend, but I can take a closer look on monday. An easy solution is to exclude some of the fcrepo-audit dependencies16:26
<awoods>acoburn: ok, I will give it a look... not that we know the source of the problem. Thanks.16:27
<acoburn>awoods: thanks
* acoburn leaves
<escowles>awoods: do we need to break out the shared code into a separate module so it doesn't have to bring all the kernel/inject baggage with it?16:40
<awoods>escowles: I don't think so...
escowles: it looks like fcrepo-audit is defining a version of javax.inject which is not the right version16:41
escowles: did you see acoburn's gist?
escowles: see: osgi.wiring.package=javax.inject)(version>=1.0.0)(!(version>=2.0.0)
* jgpawletko leaves16:42
<escowles>awoods: ok, that seems like a fixable thing
<awoods>escowles: I read that to say that fcrepo-audit expects a version 1.x of javax.inject
escowles: but the version in play is 2.3
escowles: I am trying to see where we can overwrite that dependency declaration
escowles: you can see the bad declaration if you open the fcrepo-audit.jar and look at the file: META-INF/MANIFEST.MF16:43
escowles: are you looking into how to set the version declaration correctly?16:47
<escowles>i'm still digging around the pom files and making sure everything is declaring v2.316:48
awoods: well, i'm done for the week -- have a great weekend!17:04
<awoods>escowles: ok... I will keep looking
escowles: talk to you on Monday17:05
* jcoyne leaves17:16
* whikloj leaves17:28
* barmintor leaves17:33
* dwilcox leaves17:53
* ksclarke leaves18:20
* MohamedAR leaves18:50
* pgwillia leaves18:54
* jcoyne joins19:26
* jcoyne leaves19:33
* ksclarke joins22:11
* Guest51976 leaves23:24
* ksclarke joins23:39
* ksclarke leaves

Generated by Sualtam