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

Using timezone: Eastern Standard Time
* tjohnson leaves04:56
* dwilcox joins08:47
* ajs6f joins09:00
* mikeAtUVa joins09:11
<ajs6f>awoods: I've got some Jena stuff this morning, but then I will finally get a chance to muck with the OR API presentation slides. Was there any special meaning to the order of them? They're not in correspondence with ours (that is, they're not arranged so that we have only one transition between us) so I thought I would rearrange them.09:13
<mikeAtUVa>awoods: has anyone actually tried using the backup/restore method to migrate from leveldb to mysql?
<awoods>mikeAtUVa: not that I know of09:33
mikeAtUVa: https://docs.jboss.org/author/display/MODE50/Migrating+from+3.x+and+4.x#Migratingfrom3.xand4.x-Migratingdata09:35
<mikeAtUVa>awoods: Ok... so my error may not be because of a corrupt repo (and thus flawed backup) but possibly because it just doesn't work yet... " Node types were read from the system content, and appear to be inconsistent or invalid: repo09:37
org.modeshape.jcr.value.ValueFormatException: Error converting "fedora:Resource" from String to a Name"
<awoods>ajs6f: I begin with and cover: CRUD/AuthZ/Messaging. You follow with Versioning/Fixity/Transactions/closing... Please arrange the slides accordingly.
mikeAtUVa: that looks like an exception09:38
mikeAtUVa: how did you come across it?
<mikeAtUVa>awoods: it occurs when restoring a backup to mysql. I think all the data copies and there are no mysql errors (there were until I upped some variables). The most relevant line in the stack trace is here: at org.fcrepo.http.api.repository.FedoraRepositoryRestore.runRestore(FedoraRepositoryRestore.java:84) [fcrepo-http-api-4.5.2-SNAPSHOT.jar:4.5.2-SNAPSHOT]09:40
<awoods>mikeAtUVa: can you gist the entire stacktrace?09:42
<mikeAtUVa>awoods: https://gist.github.com/mikedurbin/ed06e0eba36b8d8993236963453e33b009:43
awoods: be sure to consider the possibility the backup was made of an already corrupted leveldb database.09:44
<awoods>mikeAtUVa: can you verify if the source was corrupted?09:45
<mikeAtUVa>awood: I'm not sure how... the database was absolutely corrupted sometime after the backup.09:46
<awoods>mikeAtUVa: the error is odd: https://gist.github.com/mikedurbin/ed06e0eba36b8d8993236963453e33b0#file-gistfile1-txt-L130 , given the fact that the 'fedora' prefix is defined in the cnd: https://github.com/fcrepo4/fcrepo4/blob/master/fcrepo-kernel-modeshape/src/main/resources/fedora-node-types.cnd#L24
<mikeAtUVa>awoods: well, we're restoring to a clean (empty) repository without any registered namespaces.09:47
awood: I don't know if that's a use case we've tested or support...
<awoods>mikeAtUVa: I assume you had to at least start the new Fedora repository before restoring into it... and the cnd is loaded at start time.09:48
<mikeAtUVa>awoods: yeah, I started it... I even tried creating a resource first (I wasn't sure when that namespace was first registered).09:49
awoods: I think I should try to reproduce this in a case where I know the backup is valid... but it'll have to wait until I resolve this situtation locally.
awoods: it doesn't help that it takes 3+ hours restoring the backup before it spits out that error.09:50
<awoods>mikeAtUVa: That sounds frustrating. You could also consider the non-backup/restore pathway... that does not yet exist.09:51
<mikeAtUVa>awoods: I'm only going to the backup because the original leveldb won't start up.
<awoods>mikeAtUVa: is the production?09:52
mikeAtUVa: is this production?
<mikeAtUVa>awood: it is, but it's not public-facing, is metadata-only and redundant copies of the relevant metadata exist (in a format that might take some time to ingest).09:53
awoods: ^^^ it is inconvenient timing... but better when there's only a million resourced rather than 2M or 5M.09:54
<awoods>mikeAtUVa: do you have nightly backups? or some IT backup?09:55
<mikeAtUVa>awoods: I have some backups at key points, but haven't really considered that largely because the main activity I was performing when it got corrupted was an update to 99% of the records (to move away from federation) and it's likely less work to start fresh.09:58
awoods: especially if there's a time-consuming and untested migration from one db to another.
<awoods>mikeAtUVa: thanks for raising the issue...and sorry for the trouble.09:59
<mikeAtUVa>awoods: no worries... as I get this sorted out I'll document and test more thoroughly some of these edge cases.10:00
* peichman joins10:01
* bseeger joins10:03
* jcoyne joins10:26
* github-ff joins10:46
[fcrepo4] awoods pushed 1 new commit to master: https://git.io/voUDq
fcrepo4/master 6d73b15 Aaron Coburn: Remove all trace of javax.jcr.Node from the kernel-api (#1050)...
* github-ff leaves
* escowles joins10:52
* whikloj joins10:56
<escowles>is there a wiki page for reporting on mode5 testing?10:58
i've been running some hydra tests against it, and re-running my jmeter tests too, and i thought it would be good to post results beyond updating my jmeter test result pages10:59
<whikloj>escowles: I haven't seen one.11:00
* tjohnson joins11:04
* travis-ci joins11:06
fcrepo4/fcrepo4#4507 (master - 6d73b15 : Aaron Coburn): The build passed.
Change view : https://github.com/fcrepo4/fcrepo4/compare/2d3506051df6...6d73b15ef46f
Build details : https://travis-ci.org/fcrepo4/fcrepo4/builds/135612242
* travis-ci leaves
* barmintor joins11:11
<escowles>@whikloj: i made a page for mode5 testing results: https://wiki.duraspace.org/display/FF/Modeshape+5+Testing11:14
<awoods>escowles: I am glad to see the testing is positive so far.11:22
<escowles>awoods: yep, no problems so far - jmeter looks like it's performing better than mode4+posgres
* bseeger leaves11:30
* jcoyne leaves11:43
<ruebot>ajs6f++ #as a stand in for dan davis12:42
<ajwagner>ajs6f++ Agreed. It's great to see this "released" as we're about to roll it out in production in our "final fedora 3" infrastructure.12:48
<ajs6f>ajwagner:ruebot: Thank ddavis. He put the cash on the barrelhead for it. Actually, if you come to OR you can thank him in person.12:54
<ajwagner>I am, and I will.
<ajs6f> ajwagner: Wait, how did you even know about it before it was released?13:11
* dwilcox leaves13:13
<whikloj>illuminati confirmed...13:14
<ruebot>ajs6f: ddavis announced it on both islandora lists, and fedora-tech/community13:18
<ajs6f>ruebot: I saw that, but ajwagner said "It's great to see this "released" as we're about to roll it out in production in our "final fedora 3" infrastructure." implying that they made that decision some time ago.13:19
<whikloj>you have a mole in your organization...13:25
<ajs6f>whikloj: ddavis' org. I'm just a hired gun.13:26
<whikloj>ajs6f: perhaps you are the mole?!
ajs6f: maybe its this guy13:27
<ajs6f>whikloj: That's just me under a different ID I sometimes use as a scratch space.
<ruebot>ajs6f: i'm calling you basil from now on.13:28
<whikloj>ajs6f: but how can I know for sure...that's what the mole would say too
* bseeger joins13:29
<whikloj>ruebot: I was going to suggest Mr. Smiley
* dwilcox joins13:30
<ajs6f>whikloj: How do you know to whom you are typing right now? https://en.wikipedia.org/wiki/On_the_Internet,_nobody_knows_you're_a_dog
<whikloj>ajs6f: exactly...I assume this channel is an elaborate setup of IRC bots all trying to out-wit each other.13:31
<ruebotsbot>who am i!?
<ajs6f>whikloj: That would explain a good deal about our codebase.13:32
<whikloj>ajs6f: and the lack of conversation on voice calls
<bseeger>ajs6f, whikloj: it's hard to find a human to translate sometimes.13:33
<whikloj>agreed, I s0m3times7897989iljkd....
<ajs6f>This in-person OR meeting is going to be a sad, awkward affair. Just a room full of laptops on a conference table, softly beeping.
<whikloj>my laptop is very exciting once it has a few beers in it13:34
watch the sparks fly
<bseeger>I have Karaf running, and it could use a few beers.13:38
Hmmm… it need's a beer service.13:39
<ajwagner>ajs6f ruebot: we've been playing with trippi-sparql since April 8th.14:09
<ajs6f>ajwagner; I hope it works.14:12
<ajwagner>ajs6f: And I knew about it based on your comments here: https://jira.duraspace.org/browse/ISLANDORA-1535
It's working really well.
<ajs6f>ajwagner: Ooh, that's not good. This is Fedora. The other shoe has got to drop sooner or later14:13
<ajwagner>A single patch to islandora's tuque interface to make Fuseki happy (which wouldn't be required for rdf4j / sesame).
In testing the rebuild of the triples, one of our pids failed out of ~26,000 which I have yet to investigate, but that's the only issue so far.14:16
<whikloj>ajwagner: I'd be interested in what the patch to tuque was14:20
<barmintor>did nobody tell dan that we deprecated fcfg in favor of the Spring config14:24
sometimes I wish instead of Fedora 3.5 we had released Fedora 4, on Spring, with the object properties moved into a RELS-* datastream, and no inline XML datastreams14:26
<ajwagner>whikloj: It's hackish, and could probably done in a better way https://gist.github.com/Rarian/e809c39968bd6b243aaf1a030636d43b14:29
<whikloj>ajwagner++ # Thanks14:30
<ajwagner>I've just been adding the namespaces by monitoring fuseki logs... so I'm guessing that's not yet an exhaustive list.14:31
<ajs6f>barmintor: If you know a better way to configure the RI module… PRs welcome.14:36
barmintor: I wanted to deliver a usable product on time and under budget.
<ajwagner>ajs6f: We configure ours via spring14:38
<ajs6f>ajwagner: Let me repeat: PRs welcome!
ajwagner:barmintor: I spend just as much time with F3 as I get paid for.
<barmintor>ajs6f: I am just being ornery, but technically you *can* configure the datastores in spring14:39
<ajs6f>ajwagner:barmintor: The list of "future features" are what Dan might be paying me to work on.
<barmintor>ajs6f: and I applaud your approach to fcr3
<ajs6f>barmintor: I'm not being ornery. I'm being brutally pragmatic. This isn't about doing beautiful work. F3 is hanging off the cliff by its fingers and I am stomping on them.14:40
<barmintor>ajs6f: I have had to backtrack and rewrite a bit of my OR preso, btw, b/c I read your abstract and I think you're doing it
<ajwagner>ajs6f: barmintor: I found we had to do an ugly sed hack to make the spring config work (to delete the datastore section of fcfg)
<barmintor>ajs6f: so I am free to talk about other stuff!
<ajs6f>barmintor: Happy to talk about other stuff, but all I have to say about F3 is https://www.myinstants.com/media/sounds/nameless-scream-3-howie-scream.mp314:41
barmintor: What are awoods and I doing that you want to do? We're still writing that, more or less.14:42
<barmintor>ajs6f: I am just being much more glib about the 1998-2006 period now14:43
<ajs6f>barmintor: 2006? What happened then?14:44
<barmintor>ajs6f: "Significant growth of Fedora user community"14:45
ajs6f: http://fedorarepository.org/about/history ;)14:46
<ajs6f>barmintor: According to what sttats?
<barmintor>more seriously (I don't have my notes) I think that's when the last Fedora related article coauthored in some capacity by Payette and Lagoze was published14:47
<ajs6f>barmintor: Dusty names out of a darkening past. We scramble madly onwards towards the deadly searing light of the future.14:49
<barmintor>ajs6f: I argue that either the past is as deadly and bright as the future or that the future is as dusty and dark as the past, I can't remember which14:50
* escowles leaves
<ajs6f>barmintor: That's because although we think we're going forward, we're really spinning crazily in circles.14:51
<ajs6f>Why aren't awoods and dwilcox that inspiring?
* dwilcox leaves15:02
* dwilcox joins15:05
<ajs6f>In a vague, half-hearted way, I feel like the committers who are going to OR should probably sign keys.15:08
afk bbl
* ajs6f leaves
* bseeger leaves15:25
* github-ff joins15:28
[fcrepo-java-client] bbpennel opened pull request #12: Fluent API (master...fluent) https://git.io/voTaA
* github-ff leaves
* dwilcox leaves15:56
* bseeger joins15:59
* mikeAtUVa leaves16:05
* mikeAtUVa joins16:06
* bseeger leaves16:29
* bseeger joins16:30
* bseeger leaves16:31
* bseeger joins16:34
* ajs6f joins16:37
* bseeger leaves16:43
<ajs6f>awoods: Is the Google Hangout linked here: https://wiki.duraspace.org/display/FF/2016-06-13+-+Open+Repositories+Tech+Meeting16:45
under your Google+ account? The reason I ask is to see if we can make sure that it is a "Hangout on Air", both so that more epople can listen in if they want and so that it will get recorded.
<awoods>ajs6f: no... maybe we need a different hangout link.16:46
<ajs6f>awoods: Urb— I'm afraid I dont know much about this stuff— maybe someone at Duraspace does?16:47
<whikloj>awoods/ajs6f: http://g.uconn.edu/2014/05/16/creating-a-hangout-on-air-broadcast-and-recording/
cause I'm not getting up at 4am :P
<ajs6f>awoods: Or someone else. Maybe this kind of thing takes someone a little more Canadian than we are.
<awoods>whikloj: if you can set us up, that would be great.
<whikloj>awoods: except you won't have access to my account when you want to start your broadcast.16:48
<ajs6f>Use Fedo's account.16:49
<awoods>ajs6f: https://support.google.com/plus/answer/2553119?hl=en
<whikloj>awoods: I'll take a look and see what I get if I schedule one
<ajs6f>awoods: Are we in the videoconference room in TCD's Lloyd Building? Maybe we could use that gear...16:50
<awoods>ajs6f: is that what the schedule says?
<ajs6f>awoods: It says16:51
Location: Lloyd B
Lloyd Building (basement) - Monday only
I thought you might have more specific information.
<awoods>ajs6f: not really... although dwilcox may
<ajs6f>awoods: Eh, I think it's better to have control ourselves anyway, if we can. Maybe we can take a brief look at whatever they have that morning and see if we can make use of it.16:52
<awoods>ajs6f: if we work out the kinks of setting up a google-hangout "on air" we should be able to capture audio/video.16:53
<whikloj>awoods: Does Fedora or Duraspace have a Youtube account, that seems to be the key for the On Air feature
<ajs6f>awoods: Sure, I was thinking more of the possibility of a high-grade microphone actually calibrated for the room. That can make a huge diff in quality.16:54
<awoods>whikloj: yes, https://www.youtube.com/channel/UCE_wcxwIQCvVZkbKtC0ej3g16:55
<ajs6f>That's all boring Fedora stuff. Where are the kitten vids and pirated movies?16:56
<whikloj>awoods: well that is half the battle, do you have control or upload access to that channel?16:57
<awoods>whikloj: I am sure I could get access... but do not have it at the moment.
whikloj: I will get the details tomorrow.16:58
<ajs6f>afk bb tomorrow
* ajs6f leaves
<whikloj>awoods: we should mention this to ruebot tomorrow, he has more experience with Youtube than I. But it appears fairly straightforward17:03
awoods: https://www.youtube.com/watch?v=gJ7CtbyiLgE
<awoods>whikloj: thanks for looking into it17:04
* dwilcox joins
* mikeAtUVa leaves17:22
* peichman leaves17:23
* dwilcox leaves17:24
* dwilcox joins17:48
<ajwagner>I can bring a nicer mic to the meeting if it's wanted.17:54
* whikloj leaves18:01
<awoods>ajwagner: that would be helpful, thanks.18:13
* dwilcox leaves18:17
<ajwagner>awoods: Alrighty.21:12
* mikeAtUVa joins22:37
* mikeAtUVa leaves23:39