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

Using timezone: Eastern Standard Time
* escowles_ leaves02:06
* escowles joins02:09
* rbenji joins03:15
* rbenji leaves05:02
* kat3_drx joins08:54
* bseeger joins08:57
* rbenji joins09:00
* apb18 joins09:08
* chadmills joins09:11
* apb18 leaves09:15
* ajs6f joins09:16
* kat3_drx leaves09:30
* acoburn joins09:35
* whikloj joins09:40
* apb18 joins09:44
<whikloj>Does anyone know off-hand if the target of a ldp:membershipResource emits a JMS event when it is updated by the indirect container?09:46
<acoburn>whikloj: it doesn't, but it should and there's a ticket for that with my name on it09:47
<ajs6f>whikloj:acoburn: Did we ever come to some understanding about whether that is synch/asynch?
awoods: I will have to leave the call a bit early again today, but this will be one of the last times (maybe the last time) that will happen.09:48
<acoburn>ajs6f: the _simple_ impl of that would be sync, but it's an excellent candidate for async via akka
<ajs6f>acoburn: Oh, yeah, now I'm remembering that whole conv.09:49
<whikloj>ajs6f: I think an option would be good. For CLAW we are looking at what we want to return for pcdm:Collection info and I realized that none of the pcdm:hasMember properties are in the TS
<ajs6f>acobrn: I still wonder if it wouldn't be better to do it asynch and let synch be the single-thread limiting case.
whikloj: An option for synch/asynch? That decision shouldn't affect what eventually gets emitted— just when.09:50
<whikloj>ajs6f: yes
<ajs6f>whikloj: Everyone agrees that the containment stuff has to transmit activity and emit events. The question is just… how (as in impl)?
<whikloj>ajs6f: so (and I think this was your point) if someone wanted to take the performance hit for synch that would be their choice.09:51
<acoburn>ajs6f: the main issue I see with async relates to authZ: if a user modifies a resource that points to a ldp:membershipResource that the user is unauthorized to modify, what's the expected behavior?
<ajs6f>whikloj: Right, right.
acoburn: Oh, good point. Goodness, I hate authZ.09:52
Really hate it.
<whikloj>acoburn/ajs6f: https://jira.duraspace.org/browse/FCREPO-1889
I was trying to duplicate that last night
<ajs6f>whikloj++ # SCIENCE!
<whikloj>Except I discovered as user1 with access to some resource, I can't register a namespace :(09:53
<ajs6f>whikloj: Wait, so no one can ever use namespaces?
<whikloj>ajs6f: No, as fedoraAdmin I was fine. But as a normal user with access to a resource, I could update it if I used only prefixes and namespaces previously registered.09:54
I didn't try skipping PREFIX declarations all together and embedding the namespace in the RDF09:55
<ajs6f>whikloj: Hm. That's definitely not what I would expect, but it's interesting that no one has complained about it. Maybe it's what most people expect, or maybe no one has thought about it. Is registering a namespace prefix an 'action"?
<whikloj>ajs6f: things I will learn this morning09:56
* awoods leaves09:59
* awoods joins10:00
<whikloj>ajs6f: it appears we have the register namespace permission included in the acl:Write permission. Maybe I am doing something wrong10:03
<ajs6f>whikloj: It's a question. You are altering the state of the repo. On the other hand, you may or may not be using that facility to alter the state of objects. I would very much like to get rid of prefix registration period, and keep prefixes request or response-scoped. I don't know why we don't just do that.10:05
<whikloj>ajs6f: do you mean expand the rdf with the prefixes included before processing them?
<ajs6f>whikloj: Means a couple of things. On a request, it would mean that you need to declare the prefixes you want to use in the request. You have to do that now, so who cares?
whikloj: On a response, depends on the mimetype. If you're getting, for example, NTriples, no real change.10:07
whikloj: If you are getting an RDF serialization that understands prefixen, we could offer a way for you to request them.
whikloj: Here's the other thing we could do: actually model a namespace prefix registry in LDP, instead of the half-assed thing we have now.10:08
whikloj: What we have now basically exposes JCR functions for convenience. That's not healthy for children or other living things.10:09
whikloj: We are (correctly) moving to model versioning and auth and fixity and everything else in LDP. We could do the same with namespace prefix registries. I'm not ecstatic about that idea, but it would be better than what we have now.10:10
* dwilcox joins10:13
* apb18 leaves
<ajs6f>afk bb for the call10:14
* ajs6f leaves
* ksclarke joins10:22
* Guest84562 leaves10:29
* Guest84562 joins
* Guest84562 leaves10:30
* tjohnson joins10:34
* mcritchlow joins10:37
* jcoyne joins10:40
* jrgriffiniii joins10:49
* ajs6f joins10:56
What's the meeting page?
* yinlin joins
<acoburn>ajs6f: https://wiki.duraspace.org/display/FF/2016-04-07+-+Fedora+Tech+Meeting10:58
<ajs6f>acoburn++ # thanks
Holy cow I love that hold music.10:59
* ajs6f is here.
<bseeger>bummed I missed the music11:00
<ajs6f>bseeger: Think "a sleezy version of Hall and Oates"
<ajs6f>"Yo ho". That shold be awoods' new catchphrase.11:02
<escowles>barmintor has a star next to his name on the tech call agenda
<ajs6f>The wiki page is loading witout CSS for me. Weird.
Was it barmintor ^^^
<acoburn>* is here *
<escowles>barmintor: rotating note taking duties, but i'm guessing you're not on the call?
<barmintor>escowles: I can be in a sec
I think
yeah, meeting cancelled, one sec11:04
<ajs6f>Because Fedora is _all_ about relaxation.
Fedora 4: Smooth reggae repo vibes.
* dshalvi joins
<whikloj>ajs6f++ and ^^++
<ajs6f>I did nothing. Absoutely nothing.11:05
Maybe everyone should take notes and we merge them later.11:06
<ajs6f>Pay me money.11:07
* dwilcox leaves11:09
<barmintor>I am also editorializing a little, but you get what you get.
is there a "save and keep editing" option in confluence?11:12
<ajs6f>barmintor: Editorializing in the minutes?
<barmintor>one wonders
ajs6f: they're more notes than minutes
<ajs6f>barmintor: I thikn you can hit preview, and then you can keep hitting save, nbut I may be wrong.
<escowles>barmintor: i think it auto-saves
<ajs6f>@Ignores for tests.11:14
<ajs6f>escowles++ # can we not automated some of this stuff?11:17
<jrgriffiniii>Could this trigger a failure on the Travis slave potentially?11:19
(i. e. the installation process)
<dshalvi>we are also happy to test against MySQL if someone can point us to what is available for MySQL testing so far
* peichman joins
* ksclarke joins11:20
<ajs6f>dshalvi: I think it amounts to some manually-insertable config files.11:21
<escowles>dshalvi: ajs6f: the wiki docs for setting up mysql are here: https://wiki.duraspace.org/display/FEDORA4x/Configuring+JDBC+Object+Store
<ajs6f>WTF is the Duraspace site loading so weird for me? Is anyone else seeing this? No CSS?11:22
<escowles>ajs6f: it's working fine for me
<ajs6f>escowles: Thanks. Urb. Stupid WWW.
I would like to automate everything: coding, release, bug fixes, everything. Then I could retire.11:23
<barmintor>I would like to retire regardless of the state of automation.
<ajs6f>barmintor: I think you mean "quit".11:24
* barmintor 's head explodes
<ajs6f>barmintor: Or just "leave". "Go". "End".
"Marving K. Mooney, Will You Please Go Now."
<whikloj>Fedora 4: Less creepy11:34
<barmintor>thanks whikloj
<escowles>but not strong enough to avoid the meta joke
<barmintor>escowles: I'm only human
<ajs6f>You need metastregnth for that.
For which you need to be a metahuman.11:36
<escowles>i'm just a meathuman
<ajs6f>Leave time for staling rob.
<barmintor>+1 to memento11:37
I think the memento folks will stay silent on how POST to a TimeGate should work11:38
but we'll see
<ajs6f>barmintor: If they do, we can make a a choice, right? I'm confident that we are smart enough to make a good choice.
<barmintor>ajs6f: yes, I think it will be regarded as complementary11:39
<ajs6f>barmintor: Memento+
<whikloj>explicit +111:40
* barmintor shrugs11:42
<ajs6f>Going to have to run in a minute.11:43
Got to go11:46
<barmintor>ajs6f: ttyl!!
<ajs6f>barmintor: We need to get this versioing thing straight before Dublin, at aleast.11:48
afk bbl
* ajs6f leaves
<barmintor>we should squash all the commits for every release too11:50
single commit releases!!
<escowles>we should keep our master branch as just a single commit, and for releases we'd just force-push to master11:51
<bseeger>I'm interested, but not a committer and not completely familiar with the internals of fedora yet. Not sure I can represent.11:52
escowles: just add a line to the commit message11:53
<whikloj>bseeger: I'm an committer and not completely familiar with the internals of fedora yet. Also not sure I could represent
Also, don't use archivematica.11:54
<bseeger>whikloj - I think we may use it, or want to use it, but that might be totally separate from our fedora stuff.
<barmintor>whikloj: *you* don't use archivematica.
#fcrepo thinks archivematica is A-OK!11:55
<whikloj>barmintor: there are archivists here that might, but they don't talk to me
<barmintor>whikloj: yes, just clarifying :)
<whikloj>barmintor: clarifying that people don't talk to me?
<barmintor>whikloj: I can't answer that without undermining my goals.11:56
<barmintor>awoods: I'm turning into a pumpkin, can someone else pick up notes?11:59
<whikloj>acoburn: I could not find your ticket about emitting events from the subject of an indirect container update, if you have it I'd like to watch it.12:05
* awoods leaves
* awoods joins12:06
<acoburn>whikloj: there are a series of tickets that are all related to this: fcrepo-1498 and fcrepo-174212:08
whikloj: when those are fixed, so too will be fcrepo-188912:09
whikloj: those tickets aren't exactly about emitting events, but they amount to exactly the same thing
<whikloj>acoburn: re: fcrepo-1889 does that mean the horrible error message will go away?12:10
<acoburn>whikloj: what's the horrible error message?
* f4jenkins joins13:06
* bseeger joins13:18
* github-ff joins13:26
[fcrepo4] awoods deleted fcrepo-4.5.0-RC-1 at f6a1adc: https://git.io/vVPJ2
* github-ff leaves
* github-ff joins
[fcrepo4] awoods deleted 4.3.1-SNAPSHOT at 59defcb: https://git.io/vVPJo
* github-ff leaves
* github-ff joins
[fcrepo4] awoods deleted untagged-b124d7eb67f292743adb at fc630d0: https://git.io/vVPJX
* github-ff leaves
* github-ff joins13:27
[fcrepo4] awoods deleted 4.2.1-SNAPSHOT at 2df258d: https://git.io/vVPJM
* github-ff leaves
* github-ff joins
[fcrepo4] awoods deleted fcrepo-4.1.1-audit.1 at 2c8929d: https://git.io/vVPJ9
* github-ff leaves
<ruebot>awoods: update, we're averaging between 50-100 containers added per chunk now. no more 1000 per. resource load is really high too: https://www.dropbox.com/s/w4s1ojkp2085eaz/Screenshot%20from%202016-04-07%2013-25-00.png?dl=0
* github-ff joins
[fcrepo4] awoods deleted fcrepo-4.0.0-alpha-5 at 6c0cff3: https://git.io/vVPJ5
* github-ff leaves
* github-ff joins
[fcrepo4] awoods deleted fcrepo-4.0.0-alpha-3 at d49af01: https://git.io/vVPJp
* github-ff leaves
* github-ff joins
[fcrepo4] awoods deleted fcrepo-4.0.0-scape at 9d8a5ad: https://git.io/vVPUe
* github-ff leaves
* dwilcox joins13:28
<ruebot>awoods: https://gist.github.com/ruebot/cbfc3f89ae853dc43f899cafdd67a074 -- more info if you're curious
<awoods>ruebot: can you capture the current state, then cycle the tomcat server?13:31
ruebot: We are trying to test two things: at what point does F4 start to bog down, and how many resources can we put in an F4?13:32
* dwilcox_ joins
<ruebot>awoods: sure. ctrl+c, copy logs, restart tomcat, and fire it off again?
<awoods>ruebot: you are hitting the bog down point
ruebot: we do not yet have a good "how large can it go" number.13:33
<ruebot>awoods: cool. i'll try my best to get you that number.
* dwilcox leaves
ruebot: it will be interesting to see if F4 is snappy again after the restart.13:34
<ruebot>awoods: we are snappy again13:37
summary + 299 in 12s = 25.6/s Avg: 32 Min: 15 Max: 965 Err: 0 (0.00%) Active: 1 Started: 2 Finished: 113:38
summary + 984 in 30s = 32.8/s Avg: 27 Min: 9 Max: 617 Err: 0 (0.00%) Active: 1 Started: 2 Finished: 1
summary = 1283 in 42s = 30.8/s Avg: 28 Min: 9 Max: 965 Err: 0 (0.00%)
<awoods>ruebot: then it comes down to: what is the source of the bogging?
* dwilcox_ leaves13:39
<ruebot>awoods: yup!13:40
<f4jenkins>Project fcrepo-mint build #94: UNSTABLE in 1 min 43 sec: http://jenkins.fcrepo.org/job/fcrepo-mint/94/13:42
Project fcrepo-module-auth-rbacl build #972: UNSTABLE in 4 min 54 sec: http://jenkins.fcrepo.org/job/fcrepo-module-auth-rbacl/972/13:47
Yippee, build fixed!13:55
Project fcrepo-mint build #95: FIXED in 1 min 44 sec: http://jenkins.fcrepo.org/job/fcrepo-mint/95/
Yippee, build fixed!13:57
Project fcrepo-module-auth-rbacl build #973: FIXED in 3 min 58 sec: http://jenkins.fcrepo.org/job/fcrepo-module-auth-rbacl/973/
<awoods>acoburn/all: old release candidates and tags have been removed.13:58
awoods: https://github.com/fcrepo4/fcrepo4/releases?after=fcrepo-4.0.013:59
<awoods>acoburn: I intentionally left the beta releases.
* bseeger1 joins14:00
<acoburn>awoods: ahh, right, they aren't release candidates. That's fine
<awoods>acoburn: we can certainly discuss their removal... but they seemed fine for now.
* bseeger leaves14:01
<acoburn>awoods: I'm ok with that. plus, they're tucked away on the second page of results
* bseeger1 leaves14:02
* bseeger joins14:03
* ajs6f joins14:05
* jcoyne leaves14:08
* bseeger leaves14:29
* bseeger joins14:33
* tjohnson leaves14:34
<ruebot>awoods: test failed14:45
summary + 665 in 26.4s = 25.2/s Avg: 36 Min: 11 Max: 802 Err: 1 (0.15%) Active: 0 Started: 2 Finished: 2
summary = 118288 in 3788s = 31.2/s Avg: 28 Min: 7 Max: 36982 Err: 1 (0.00%)
awoods: i got nothing in catalina.out error-wise, but high load, cpu usage and memory usage now. we're right around 10 million containers
<awoods>ruebot: how are you keeping container count?14:47
<ruebot>adding up the numbers at the end of each summary log14:48
awoods: i could do a query in fcrepo too... 9,153,862 is my number now14:49
* ajs6f leaves14:50
<awoods>ruebot: query?
ruebot: when you say query, do you mean ask for the root and see how many children are there?
<ruebot>awoods: heh. yes.14:51
<awoods>ruebot: https://jira.duraspace.org/browse/FCREPO-188014:52
* ruebot watches14:53
* ajs6f joins15:04
* bseeger leaves
* bseeger joins15:06
* jcoyne joins15:07
* jcoyne leaves15:18
* rbenji leaves15:34
<ajs6f>awoods: What do you think about a "mini-sprint" for the versioning spec? This is the kind of thing for which traditionally we have used a special topics call.15:36
awoods: But i think we need that _and_ more.
awoods: A call is good for those people who already know they want to engage in the conv.
awoods: But not others.
<awoods>ajs6f: I think spec sprints are a great idea15:37
<ajs6f>awoods: Sure, but in what do they consist?15:38
* chadmills leaves15:49
* dwilcox joins15:57
<ajs6f>Can I just say how awesome I think it is that Scott Prater got a response that was the single word, "false"?16:12
He asked the repo for something, and the reply was a simple, flat, "false". No further explanation. Now _that's_ the Fedora API
acoburn: Did you ever look at https://manifoldcf.apache.org/? I mean, I'm not interested in redeplying our investment in Camel, but that's kind of interesting.16:16
<acoburn>ajs6f: I've only given it a cursory glance. The documentation leaves something to be desired16:18
ajs6f: seems their focus in on sharepoint and such things
<ajs6f>acoburn: Yes, I agree. Just something to be aware of.
* rbenji joins16:25
<ajs6f>awoods: Do you want to start an OR page, since notices are out? Do we have one I can copy?
<awoods>ajs6f: dwilcox may have an OR page from the past for you... if you have not already found one to copy.16:27
<ajs6f>awoods: I can't even try: none of the Duraspace wiki pages load properly for me today.
<ajs6f>awoods: I feel like Duraspace is saying hate to me.
<awoods>ajs6f: here is a little duraspace love: https://wiki.duraspace.org/display/FF/2015-06+Open+Repositories
<ajs6f>awoods: Thanks. That doesn't load for me either, bt maybe it will tomorrow. Welcome to the paperless office of the THE FUTURE!16:32
* dwilcox leaves16:42
<ruebot>awoods: curl -i http://localhost:8080/fcrepo/rest -- has been running for over an hour :-)16:45
<awoods>ruebot: hence the ticket I posted
<ajs6f>out for the day, bye bye!
<awoods>ruebot: with millions of resources, it will run for days
* ajs6f leaves
<ruebot>awoods: good to know!16:47
* jcoyne joins
* bseeger leaves16:51
* dwilcox joins16:54
* dshalvi leaves17:16
* acoburn leaves17:18
* rbenji leaves17:19
* dwilcox leaves
* dwilcox joins17:21
* github-ff joins17:31
[fcrepo-camel-toolbox] awoods pushed 1 new commit to master: https://git.io/vVPbi
fcrepo-camel-toolbox/master a1a92c8 Aaron Coburn: Update reindexing code to properly use the camel-jetty component...
* github-ff leaves
* github-ff joins17:32
[fcrepo-camel-toolbox] awoods closed pull request #77: Fix broken reindexing (master...fcrepo-1980) https://git.io/vViSN
* github-ff leaves
* peichman leaves17:38
* dwilcox leaves17:40
* dwilcox joins17:47
* jrgriffiniii leaves17:53
* dwilcox leaves17:54
* travis-ci joins
fcrepo4-exts/fcrepo-camel-toolbox#219 (master - a1a92c8 : Aaron Coburn): The build was broken.
Change view : https://github.com/fcrepo4-exts/fcrepo-camel-toolbox/compare/91106baea477...a1a92c87e336
Build details : https://travis-ci.org/fcrepo4-exts/fcrepo-camel-toolbox/builds/121548347
* travis-ci leaves
* whikloj leaves17:57
* mcritchlow leaves18:00
* jrgriffiniii joins18:52
* jcoyne leaves19:01
<awoods>barmintor: the migration-utils has a critical bug19:12
* ksclarke leaves19:55
* jgpawletko joins20:07
* jgpawletko leaves20:36
* peichman joins21:24
* peichman leaves21:25
* jrgriffiniii leaves22:16
* jcoyne joins22:44
* dwilcox joins23:05
* jcoyne leaves00:04
* jrgriffiniii joins00:51
* dwilcox leaves01:04
* dwilcox joins01:12
* dwilcox leaves01:26