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

Using timezone: Eastern Standard Time
* dwilcox joins08:10
* awead joins08:23
* acoburn joins08:46
* mikeAtUVa joins08:53
* ksclarke joins09:39
* jgpawletko joins10:18
* longshou joins10:45
* scossu joins10:51
* scossu leaves10:57
* JohnDoyle_NLM joins11:07
* dshalvi joins11:11
* escowles joins
* awead_ joins
* Yinlin joins11:13
* Yinlin leaves11:17
* awead leaves11:18
* scossu joins11:27
* scossu leaves11:36
<escowles>awoods: i'm dropping off the call -- my $0.02 on tech meetings for the rest of the year: i'll be around 12/11 & 12/1811:58
* JohnDoyle_NLM leaves12:03
<ruebot>i'm dropping now. cheers!12:04
<dshalvi>Thanks all for releasing Fedora 3.8 and 4.0!
* github-ff joins12:19
[fcrepo-camel] acoburn opened pull request #21: osgi deployment requires that getters and setters use the same types (master...osgi-deployment-support) http://git.io/XduyhA
* github-ff leaves
<awoods>escowles: ping12:21
* scossu joins12:22
<awoods>mikeAtUVa: we missed you on the tech call
* dshalvi leaves12:26
* scossu leaves12:30
<ksclarke>awoods: hmm, I added a note on a jira ticket that included [~awoods] but it doesn't seem to have translated that into a link like Jira usually does... do you see a link (or did you get notice of it)? https://jira.duraspace.org/browse/FCREPO-123512:44
<awoods>ksclarke: I did get a notice... but not because of the reference to my name (I get noticed of all updates in the FCREPO JIRA project).12:45
<ksclarke>awoods: shouldn't that have translated into a link (and an additional alert be raised) or is that a function of a plugin that's not installed/configured or something?12:46
It did let me search your name, and it input the [~awoods] form for me when I selected your name from the dropdown
<awoods>ksclarke: it should. I will look into why those name-references are not being properly interpreted.
* longshou leaves12:58
* scossu joins13:16
<awoods>ksclarke: can you please add your meeting notes to the wiki when you have a chance?13:25
<ksclarke>yes, was sidetracked a bit... will do though
awoods ^
* scossu leaves
<awoods>ksclarke: thanks
* longshou joins13:34
<awoods>longshou: jira avatar?13:35
<escowles>awoods: yes?13:36
<awoods>escowles: hi
<escowles>awoods: i asked matt to send you a diagram of our JIRA workflow13:37
<awoods>escowles: I wanted to ask you about the OOM/too-many-threads issue that you have seen (and resolved) in the past.
escowles: do you have the F4 code in front of you?
escowles: I am currently upgrading from ModeShape 4.0.0 to 4.1.0...13:38
<escowles>i haven't updated in a couple weeks, let me pull master...
<awoods>escowles: ModeShape has introduced a minor API change...
escowles: which relates to the issue of the ObservationManager and registering (and not unregistering)13:39
escowles: It looks like the "fix/workaround" that we have in place is no longer needed...
<escowles>awoods: so related to the event generation?
<awoods>escowles: yes
escowles: I was wondering if you could test/reproduce the scenario that surfaced that error.13:40
escowles: I thought you and jcoyne saw the issue while performing simple ingests.
escowles: do you have the code up?
<escowles>awoods: yes, i think it would be pretty easy to test -- we could just remove the code that unregisters and then do some simple load test13:41
awoods: i've got my code updated -- we are talking about fcrepo-http-api, right?
<awoods>escowles: I am running a test and my machine is a little slow at the moment, sorry.13:42
escowles: see SessionProvider.dispose()13:43
escowles: which calls SessionUtils.unregisterObservationManager()
<escowles>awoods: yes, i think you could simply remove the SessionProvider.dispose() method and then make 1000 curl requests to reliably trigger the issue13:44
<awoods>escowles: if you update your master pom.xml to use:13:45
- <modeshape.version>4.0.0.Final</modeshape.version>
+ <modeshape.version>4.1.0.Final</modeshape.version>
escowles: you will see that SessionUtils now has a compilation error (due to the ModeShape change)13:47
escowles: but given the recent ModeShape update, our workaround may no longer be necessary: https://github.com/ModeShape/modeshape/commit/8c63c406708186a644065c951d6931708bf11d63#diff-8849dd4607bd6c07a73944d2617a5e2fL17413:48
<escowles>awoods: ok, i updated to mode 4.1.0, removed the call to SessionUtils.unregisterObservationManager(), and ran my 1000 request script and didn't see any errors14:01
let me try removing SessionUtils.unregisterObservationManager() call with mode 4.0.0 and see if my script produces the OOM error
* scossu joins14:11
<awoods>escowles: I just did a Modeshape-4.1.0 ingest of 43,000 containers... no OOM.14:12
escowles: It would be great if you can show the issue in MS-4.0.0 and not in MS-4.1.0
<escowles>awoods: i just finished running my tests -- i can trigger the OOM error with mode 4.0 without the unregistration, but it works fine with mode 4.1
<awoods>escowles: perfect. It looks like we can remove the SessionUtils code.14:13
<escowles>awoods: sounds good -- do you want me to push my mode 4.1 branch i was using to test this?14:14
<awoods>escowles: yes, if you can also include in that branch the removal of SessionUtils.14:15
escowles: all you: https://jira.duraspace.org/browse/FCREPO-1249
<escowles>awoods: i didn't remove SessionUtils completely -- i just removed the unregister call14:16
<awoods>escowles: there is nothing else in that class
<escowles>awoods: oh, now i see the rest was private anyway... -- i'll remove them and then put in a PR14:17
<awoods>escowles: thanks!14:18
* scossu leaves14:20
* awoods leaves14:35
* github-ff joins14:42
[fcrepo4] escowles created mode-4.1.0 (+3 new commits): http://git.io/DDffTg
fcrepo4/mode-4.1.0 123472e Esmé Cowles: Removing manual observation unregistration
fcrepo4/mode-4.1.0 1d4a91b Esmé Cowles: Removing commented-out lines
fcrepo4/mode-4.1.0 7a29c4e Esmé Cowles: Marking failing tests pending
* github-ff leaves
* github-ff joins14:44
[fcrepo4] escowles opened pull request #672: Upgrading to ModeShape 4.1.0 (master...mode-4.1.0) http://git.io/OoEOWA
* github-ff leaves
* scossu joins15:05
* scossu leaves15:14
* scossu joins15:39
* ksclarke leaves16:00
* dwilcox leaves16:01
* ksclarke joins
* awoods joins16:47
* mikeAtUVa leaves16:58
* acoburn leaves17:40
* ksclarke leaves18:08
* awead leaves18:36
* awead joins18:39
* scossu1 joins18:44
* scossu1 leaves18:45
* scossu leaves18:48
* awead leaves18:50
* awead joins18:51
* github-ff joins19:03
[fcrepo4] awoods pushed 1 new commit to master: http://git.io/GDK_0A
fcrepo4/master ef5cc86 Esmé Cowles: Upgrade ModeShape to version 4.1.0.Final...
* github-ff leaves
* github-ff joins19:04
[fcrepo4] awoods deleted mode-4.1.0 at 7a29c4e: http://git.io/lVIcnA
* github-ff leaves
* github-ff joins19:10
[fcrepo4-oaiprovider] lsitu opened pull request #10: Added support to allow admins to manage descriptive content about the repository and make the content available. (master...feature/descriptive-content) http://git.io/3qT6ZA
* github-ff leaves
* awead leaves19:11
<f4jenkins>Project fcrepo-module-auth-rbacl build #429: UNSTABLE in 3 min 26 sec: http://jenkins.fcrepo.org/job/fcrepo-module-auth-rbacl/429/19:19
* travis-ci joins19:20
fcrepo4/fcrepo4#3252 (master - ef5cc86 : Esmé Cowles): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/57090df89204...ef5cc864705a
Build details : http://travis-ci.org/fcrepo4/fcrepo4/builds/43046628
* travis-ci leaves
<f4jenkins>Yippee, build fixed!19:34
Project fcrepo-module-auth-rbacl build #430: FIXED in 3 min 53 sec: http://jenkins.fcrepo.org/job/fcrepo-module-auth-rbacl/430/
* awead joins20:39
* longshou leaves21:09
* ksclarke joins21:29
* jgpawletko leaves23:14
* awead leaves23:23
* longshou joins23:27
* longshou leaves23:42
* scossu joins23:45
* scossu leaves23:53
* awead joins23:59
* ksclarke leaves00:29

Generated by Sualtam