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

Using timezone: Eastern Standard Time
* justinsimpson leaves01:34
* dwilcox joins01:47
* dwilcox leaves01:56
* thomz joins02:29
* dwilcox joins02:58
* dwilcox leaves04:23
* dwilcox joins04:57
* dwilcox leaves05:08
* dwilcox joins
* dwilcox leaves06:24
* youn joins06:56
* youn leaves
* dwilcox joins07:02
* youn joins07:15
* jjtuttle leaves07:32
* youn leaves07:34
* jjtuttle joins07:39
* youn joins07:44
* youn leaves07:58
* youn joins08:05
* dwilcox leaves08:15
* coblej joins08:17
* youn leaves08:19
* manez joins08:27
* dhlamb joins08:32
* westgard joins08:36
* youn joins
[Import/Export Standup]08:38
* Completed yesterday: scripts to process and load lubm data
* Planning on completing today: sparql queries; help with test plan if needed
* Blockers / Need help with: load script halts (performance issue?)
<westgard>youn: I can try out your load script if that helps. Where can I access it?08:40
<youn>westgard: thanks! here: https://drive.google.com/drive/folders/0B42p4oXygqHfS2J6MmRoNEQ4WEU08:45
<westgard>youn: Cool -- I'll take a look. I need to sort out my pull request first, but should get to this shortly.08:46
* acoburn joins08:55
* whikloj joins09:06
* bseeger joins09:15
* mikeAtUVa joins
<bseeger>[Import / Export Standup]09:18
* Completed Yesterday: meeting
* Planning on completing today:
- add specifics about tests to tests plans; load test data if possible.
* Blockers / Need help with:
- none
* ajs6f joins09:23
* manez leaves
* peichman joins09:29
<ajs6f>awoods: Do you know if scossu will be at this morning's meeting, re: FCREPO-1928?09:31
<awoods>ajs6f: I do not.
* acoburn leaves09:33
<ajs6f>awoods: Hm. I thikn I can reason pretty well about it, but I have two assumptions about his use case that merit verification by himself.
<awoods>ajs6f: have you reached out to Stefano?
<ajs6f>awoods: No, because I didn't know the issue existed until you sent out a message this morning.09:34
awoods: And communicating with users is way more effort than I am going to put into Fedora.
<escowles>[import/export standup]:09:35
* yesterday: finished basic import
* today: translating URIs
* blockers: none
* dchandekstark joins
<ruebot>[Import / Export Standup]09:38
* Completed Yesterday:
- Tested, reviewed, merged https://github.com/fcrepo4-labs/fcrepo-import-export/pull/21
* Planning on completing today:
- https://jira.duraspace.org/browse/FCREPO-2178
- Review test https://github.com/fcrepo4-labs/fcrepo-import-export/pull/25
- Take Youn's test data for a spin
* Blockers / Need help with:
- My schedule :-(
<awoods>[Import/Export Standup]09:42
* Completed yesterday:
- Remove URISyntaxExceptions: https://jira.duraspace.org/browse/FCREPO-2183
- Add support for config file: https://jira.duraspace.org/browse/FCREPO-2164
* Planning on completing today:
- Response to code review comments on FCREPO-216409:43
- Testing datasets
* Blockers / Need help with:
- None
* justinsimpson joins09:44
[Import/Export Standup]
Yesterday:
no progress, too many meetings
Today:
starting with https://jira.duraspace.org/browse/FCREPO-2143
planning to finish #2143, #2144, #2145
Blockers/Questions:09:45
none
* youn leaves09:48
<bseeger>if I'm doing a fcr:backup and fcr:restore, should I assume that the createdAt and lastModified dates would be maintained?09:51
<awoods>bseeger: yes09:52
<westgard>[Import/Export Standup]
* completed yesterday/this morning:
- pull request that should resolve FCREPO-2166, 67, & 68
* planning on completing today:
- try out youn's data and import shell script
- verify and if necessary file a bug report for an issue discovered yesterday by which a typo in SPARQL update can cause binaries to disappear (404 not found though they are in fact still there)
* dchandekstark leaves
<westgard>* blockers:
- none
* github-ff joins09:53
[fcrepo-import-export] mikedurbin opened pull request #26: Added authorization support for import/export. (master...FCREPO-2163) https://git.io/vilzu
* github-ff leaves
<bseeger>awoods: thanks. I upgraded my 4.6 system to 4.7 snapshot yesterday (using fs for modeshape, no db) and I don't think the dates were maintained, but I'm parsing the document.bin file and it's hard to tell.09:54
<awoods>bseeger: Is there a reason you are not just looking directly at the restored Fedora resources?09:56
<bseeger>awoods: yes, I'm totally doing that. And then looking at the bin doc to verify them.09:57
* coblej leaves
<bseeger>awoods: the dates at least, since I saw an instance where the root node didn't have createdAt and lastModified anymore.
<mikeAtUVa>[Import/Export Standup]09:58
* Completed yesterday:
- https://github.com/fcrepo4-labs/fcrepo-import-export/pull/17
* Planning on completing today:
- https://jira.duraspace.org/browse/FCREPO-2163
- https://jira.duraspace.org/browse/FCREPO-2181
* Blockers / Need help with:
- None
<bseeger>awoods: I also upgraded another system to using the 4.7 snapshot, this one with a postgres db in it - that system didn't show the createdAt and lastModified at dates anymore. Just got me wondering.
awoods: to clarify - this was for the root node only - the rest seem fine09:59
* youn joins
<awoods>bseeger: It is helpful to know you are just referring to the root.10:00
* coblej joins10:01
<awoods>bseeger: The restore operation is a pure import of the previous data. I would expect even the root to be the same as before... but the proof is in what actually happens.
<escowles>bseeger: do other properties on the root work? e.g., if you set a title or other properties?10:02
* dchandekstark joins10:03
<bseeger>awoods: Okay. I need to run some more tests because to verify more of what I think I'm seeing.
escowles: yes, they seem to. I had a dc:title and skos:prefLabel set and they came through just fine.
<escowles>bseeger: ok, that's good — very strange about createdAt and lastModified not showing up, but we can't set those via the REST API anyway, so i'm not sure why that would happen10:04
* manez joins10:24
<bseeger>import/export folks - can someone remind me of where the _google_ test data is?
<escowles>bseeger: https://drive.google.com/drive/folders/0B42p4oXygqHfS2J6MmRoNEQ4WEU10:25
<bseeger>escowles++
<youn>bseeger++ escowles++ :)10:26
<ruebot>bseeger: it's on the sprint resources page too :-)10:27
* manez leaves10:28
<bseeger>ruebot: thanks - I just found that. :)10:29
* justinsimpson leaves
<bseeger>ruebot: or rather, I didn't know it existed.
<ruebot_away>awoods: i'll be away for the tech call. doing interviews.10:37
<awoods>ruebot_away: thanks, maybe we can circle up later.
* dchandekstark leaves10:39
* dhlamb leaves10:40
* dchandekstark joins10:44
* dhlamb joins10:45
* coblej leaves10:51
* coblej joins10:52
* dchandekstark leaves10:55
* coblej leaves10:57
* dchandekstark joins
* youn leaves
* dchandekstark leaves
* jenlindner joins10:58
<whikloj>https://wiki.duraspace.org/display/FF/2016-09-08+-+Fedora+Tech+Meeting10:59
<bseeger>* is here *11:00
* dchandekstark joins
* coblej joins
* apb18 joins11:02
* escowles is here
* ajs6f is here
* dhlamb is here11:03
lol
<jenlindner>lol
* thomz leaves
<whikloj>"Call 1-800-DOCTORB. The extra B is for Bargain"
* dchandekstark leaves11:05
<jenlindner>"archivematica brew bags" ?11:08
is that correct?
<escowles>jenlindner: i thought he said that archivematica produces bags
* manez joins11:11
<jenlindner>okay thanks
thanks esme11:12
bell's two hearted!
<ajs6f>Founder's beers are also pretty awesome close to the brewery, too.
Although B2H was really a revelation. All these high-register floral notes in the hops, that I had never tasted before.11:13
<westgard>This brings back many fond memories of afternoons spent in the Bell's beer garden in my Kzoo days. :-)11:16
<escowles>btw, that PR is https://github.com/projecthydra/curation_concerns/pull/90111:20
<jenlindner>who is this discussing stefano's use case?11:24
<escowles>jenlindner: adam soroka
<whikloj>PATCH11:25
<dhlamb>ajs6f: yeah, i'd say delta11:26
* dwilcox joins11:27
<jenlindner>thanks
what are the more convenient interfaces to patch?
<escowles>i think ldpatch is the main alternative i've heard of (https://www.w3.org/TR/ldpatch/)11:28
<dhlamb>ajs6f: fair enough
<ajs6f>jenlindner: RDF Patch, LDP Patch
<escowles>though regardless of the format, it's a pain to have to do a diff to generate the sparql update or ldpatch update
<ajs6f>https://dvcs.w3.org/hg/ldpwg/raw-file/ldpatch/ldpatch.html
escowles: You needn't do a diff with Rdf Patch11:29
https://afs.github.io/rdf-patch/
<dhlamb>in order to diff, you need canonical representation
<escowles>ajs6f: that sounds like a nice feature...
<ajs6f>https://www.w3.org/TR/ldpatch/
<jenlindner>thanks
<ajs6f>escowles: Andy Seaborne is nobody's fool.
dhlamb: RDF diff is pretty easy— EXCEPT BLANK NODES. **(&^*)^)*^ blank nodes.11:30
" I want it to be this state now" is definitely PUT.11:31
<whikloj>apb18++11:32
* github-ff joins11:35
[fcrepo-import-export] escowles created uri-translation (+1 new commit): https://git.io/vilSh
fcrepo-import-export/uri-translation e2c6490 Esmé Cowles: Translating URIs on import
* github-ff leaves
* github-ff joins11:37
[fcrepo-import-export] escowles opened pull request #27: Translating URIs on import (master...uri-translation) https://git.io/vil9n
* github-ff leaves
<jenlindner>what' s the header u send now ? lenient headers to not send sm triples?
<escowles>jenlindner: Prefer: handling=lenient; received="minimal"11:38
<whikloj>escowles++
<jenlindner>thanks
<ajs6f>I am conceptualizing the heck out it RIGHT NOW.11:45
* ajs6f conceptualizes like a hurricane
<jenlindner>who ssaid idempotent is not concerned with responses?11:48
<whikloj>jenlindner: I think that was aaron birkland
<ajs6f>I agree with what aaron said about that.11:49
<jenlindner>thanks
<whikloj>"From a RESTful service standpoint, for an operation (or service call) to be idempotent, clients can make that same call repeatedly while producing the same result. In other words, making multiple identical requests has the same effect as making a single request. Note that while idempotent operations produce the same result on the server (no side effects), the response itself may not be the same (e.g. a resource's state may change be
tween requests)."
<ajs6f>The only way responses come up is in the sense that in order to verify that state is what you think it is, you look at responses.
whikloj: How does "idempotent operations produce the same result on the server" work against "a resource's state may change between requests"?11:50
whikloj: We have to unpack "the same result".
whikloj: Does that mean: "initiates the same change (in the sense of delta)" or does it mean "moves the server to the same state"?11:51
<jenlindner>lol
:)
<ajs6f>+111:52
<whikloj>ajs6f: I'd say initiates the change, or attempts to but of course in the case for a DELETE you might only be able to delete it once
escowles++
<dhlamb>the "wait, what?" moment11:53
<ajs6f>whikloj: Okay, that's fine. And DELETE works in that sense because you can initiate the same change fifteen times and just because only the first completes successfully doesn't mean that the attempted changes weren't all equivalent.
<whikloj>ajs6f: exactly, that is how I read that explanation
<ajs6f>whikloj: So that definitely supports apb18's understanding of idempotency.11:54
whikloj: And it seems to me to mean that our PUT is actually okay right now.
<whikloj>ajs6f: agreed, also diegopino prefers the current functionality11:55
* github-ff joins
[fcrepo-import-export] mikedurbin opened pull request #28: Updated path escaping to avoid ambiguity. (master...FCREPO-2181) https://git.io/vil7a
* github-ff leaves
* ajs6f already released it. DONE!
<jenlindner>haha11:56
* dchandekstark joins
<ajs6f>mikeAtUVa++
I am a little cautious about using a full-on scripting language for build defs.11:59
* peichman leaves
<whikloj>GOODBYE ALL12:01
<jenlindner>lol
<dhlamb>bye everyone
* dchandekstark leaves
<ajs6f>https://www.youtube.com/watch?v=Arf3HC2S7D8
<bseeger>whikloj: thanks for adding that clarification - what you said exactly (regarding gradle and fcrepo-exts)
<whikloj>bseeger: no problem12:02
* coblej leaves12:03
* jenlindner leaves12:04
<ajs6f>awoods: You have been working on that ticket 1928, do you want to make the response to scossu?12:05
<bseeger>ajs6f: why cautious about a full-on scripting language for build defs?
* awoods on a call
but, yes
<ajs6f>bseeger: The worst thing that can happen with Maven is generally: your build fails. The worst thing that can happen due to problems in an arbitrary script is: arbitrarily bad. Not a total objection, just a root of caution.12:06
* jenlindner joins12:07
<bseeger>ajs6f: that's a good point. I wonder what boundaries are imposed in it, if any, so that you can't execute something that hurts your system/code/whatnot.12:08
<whikloj>ajs6f: fair point, but I think even if you switch to using Gradle without the scripting. It is (IMHO) much clearer to read
but I hate maven
<ajs6f>I'm not against Gradle. I am interested to hear about the advantages other than concision.12:10
<mikeAtUVa>So we're switching to ANT?
<ajs6f>I don't have a problem reading Maven
* ajs6f slaps mikeAtUVa with a trout12:11
<bseeger>apparently it's a bit faster, so I hear.
<mikeAtUVa>ajs6f: in homage to mIRC?
<ajs6f>mikeAtUVa: This http://www.mirc.com/ ? I'm afraid I don't know what you mean...12:12
<whikloj>mikeAtUVa: I have an old indexer that still uses ANT... it works well.. enough
<mikeAtUVa>ajs6f: since 1995 that IRC client had a context menu option called "/slap" that ouput that text about the trout.12:14
<whikloj>ajs6f: https://en.wikipedia.org/wiki/Wikipedia:Whacking_with_a_wet_trout12:15
<ajs6f>mikeAtUVa:whikloj: I think of https://www.youtube.com/watch?v=GSTGMwywBtc but maybe I am showing my age.12:16
<whikloj>ajs6f++
<mikeAtUVa>That was probably his inspiration...12:17
<ajs6f>afk bbl12:19
* ajs6f leaves
<jenlindner>another note-taker question, what are the names of the two java devs who will have time to address fedora performance issues? i only got danny and bill12:20
<mikeAtUVa>jenlindner: Bill Branan and Danny Bernstein12:21
<jenlindner>thanks v much
oh it's in the agenda, too12:22
* bseeger leaves12:29
* github-ff joins12:51
[fcrepo-import-export] escowles pushed 1 new commit to uri-translation: https://git.io/viljh
fcrepo-import-export/uri-translation c0c1fc9 Esmé Cowles: Addressing code review comments
* github-ff leaves
* dchandekstark joins13:00
* jenlindner leaves13:03
* tpendragon leaves13:05
* dwilcox leaves13:07
* 7GHABY88F joins13:09
[fcrepo-import-export] escowles pushed 2 new commits to master: https://git.io/vi8Jf
fcrepo-import-export/master 3637bce Mike Durbin: Updated path escaping to avoid ambiguity.
fcrepo-import-export/master e3079d2 Esmé Cowles: Merge pull request #28 from mikedurbin/FCREPO-2181...
* 7GHABY88F leaves
* 17SAA33PX joins
[fcrepo-import-export] escowles closed pull request #28: Updated path escaping to avoid ambiguity. (master...FCREPO-2181) https://git.io/vil7a
* 17SAA33PX leaves
* dchandekstark leaves13:10
* bseeger joins13:12
* dchandekstark joins13:14
* justinsimpson joins13:19
* dchandekstark leaves13:28
* youn joins13:31
* dchandekstark joins13:32
* dchandekstark leaves13:35
* dchandekstark joins13:36
* dchandekstark leaves13:39
* dchandekstark joins13:43
* tpendragon joins
* apb18 leaves13:45
* youn leaves13:47
* dwilcox joins
* dchandekstark leaves13:49
* youn joins13:51
* acoburn joins13:57
* dchandekstark joins13:58
* bseeger leaves14:02
* dchandekstark leaves14:04
* tpendragon leaves14:05
* dchandekstark joins14:07
* tpendragon joins14:08
* dchandekstark leaves14:09
* dchandekstark joins14:15
* dchandekstark leaves14:17
* dchandekstark joins14:19
* dchandekstark leaves14:20
* ajs6f joins14:22
* dchandek_ joins14:26
* dchandek_ leaves14:30
* justinsimpson leaves14:34
* github-ff joins
[fcrepo-import-export] escowles pushed 1 new commit to master: https://git.io/vi8G7
fcrepo-import-export/master 7eafab0 Andrew Woods: Add support for config file (#25)...
* github-ff leaves
<acoburn>ajs6f: so I hear we'll be migrating the fcrepo build machinery to gradle?
<whikloj>acoburn++14:35
<acoburn>whikloj: I wasn't at the meeting — it's just a rumor I heard
<awoods>acoburn: rumor14:36
<ajs6f>acoburn: Almost. Actually, we're going to rewrite Maven in Gradle and use that.
<whikloj>acoburn: totally a rumour...very bad rumour
* apb18 joins14:38
* dbernstein joins14:46
* dchandekstark joins14:59
<awoods>escowles: is this ready for review? https://jira.duraspace.org/browse/FCREPO-218015:02
* dchandekstark leaves15:03
* youn leaves15:04
* cm_harlow joins
<escowles>awoods: not quite — i need to figure out how to close the Triple Sink propertly...15:05
<awoods>escowles: just checking
* cm_harlow leaves15:08
* youn joins15:14
* github-ff joins15:16
[fcrepo-import-export] escowles pushed 1 new commit to uri-translation: https://git.io/vi8BR
fcrepo-import-export/uri-translation cbf22cf Esmé Cowles: Flushing Triple Sink, adding Javadoc comment
* github-ff leaves
* dchandekstark joins15:17
<mikeAtUVa>awoods, escowles: quick question... in rebasing the config files stuff into the authoization stuff: I don't think we should write people's password to a file. How do you feel?15:20
<escowles>mikeAtUVa: i agree it seems like a bad thing to do — but you can't mix and match command line with config file now, so that would mean you couldn't use the config file with auth...15:21
<awoods>mikeAtUVa: you could just add "***"
escowles: true, the user would have to re-add their username/password to the config file.15:22
<mikeAtUVa>awoods, escowles: as for mixing and matching... I think I'd like to have it load configs from the file if specified, then overwrite/add any options that are on the command line. Then I'd specifically omit the username from the file that's written and we'd be good, right?
<escowles>awoods: i bet most people would be fine with doing that, but writing it to a file in /tmp does seem like too permissive for some
mikeAtUVa: that sounds good to me15:23
* jam265 joins15:25
<westgard>youn: I'm testing the lubm dataset loader and I think I'm seeing the same issue you had reported. The loader shell script is continuing, but seems to be getting slower and slower. Activity monitor shows VBox using all of its allocated memory (other applications are not affected).
<awoods>mikeAtUVa: it becomes more complicated if you choose to merge commandline args and the config file... but it is certainly possible.
* cm_harlow joins15:26
<awoods>mikeAtUVa: the other option is to not write the username/password in the saved config file, and users can re-add those credentials to the file if they like.
<mikeAtUVa>awoods: if I can't work out the merge in a few minutes I'll fall back to that option.15:27
<awoods>mikeAtUVa: good luck, and be safe.
<youn>westgard: thanks for letting me. I realized I forgot to include ub types with the sparql updates, so I'm trying to capture that plus other ldp relationships. Given slowness with creating the nodes, I may try to consolidate the sparql updates but maybe not in time for this sprint.15:29
<westgard>youn: that makes sense. I think it would be great to have this sort of data available as part of a test suite, but given how long the current loader is taking, it may not be practical for testing purposes. Maybe there is a way for us to speed up the load process or reduce the size of the dataset? Some things to think about anyway.15:32
<amccarty>Hey folks - I'm new to fedora but am helping some folks here who are having an issue with possible data corruption under 4.5.115:35
<escowles>amccarty: we've seen some corruption with leveldb, particularly when the JVM hangs or is killed15:36
<amccarty>I'm at a bit of a standstill as I'm getting the following error (and tomcat fails to start the app):
org.springframework.beans.factory.BeanCreationException: Error creating bean with name 'org.fcrepo.kernel.modeshape.observer.SimpleObserver#0': Invocation of init method failed; nested exception is org.modeshape.jcr.cache.NodeNotFoundInParentException: Cannot locate child node: 87a0a8c317f1e7jcr:system within parent: 87a0a8c7505d64/
escowles: well, at least I'm not alone :)
<escowles>amccarty: cold comfort, i know15:37
but there are some scripts that have fixed some of the errors, let me see if i can find them for you
<ajs6f> amccarty: Hm. Can you show us a complete stack trace?
amccarty: Maybe gist it?15:38
<amccarty>ajs6f: yep one sec
* youn leaves
<amccarty>https://gist.github.com/amccarty/f398a1f94dcf4c6e49a76f3905bff2d415:40
more accurately: https://gist.github.com/amccarty/f398a1f94dcf4c6e49a76f3905bff2d4#file-catalina-out-L89515:41
<escowles>amccarty: the scripts i was thinking of are here: https://wiki.duraspace.org/display/FEDORA4x/Backup+and+Restore#BackupandRestore-repair-leveldb15:42
* westgard leaves
<escowles>that page has some notes on backing up and verifying leveldb files too
<ajs6f>That looks slightly odd.
escowles:dchandekstark:jjtuttle: Is that the kind of thing you saw? I find it odd that the repo doesn't crash out until the observer tries to log in.
<awoods>ajs6f/amccarty: yes, the "Cannot locate child node: 87a0a8c317f1e7jcr:system within parent" message looks similar to what Duke was seeing.15:43
<dchandekstark>ajs6f: sorry, i'm missing some context
<ajs6f>dchandekstark: It's cool, it seems that awoods answered for you.
<escowles>dchandekstark: amccarty posted a gist of an error he's seeing with leveldb: https://gist.github.com/amccarty/f398a1f94dcf4c6e49a76f3905bff2d4#file-catalina-out-L895
<ajs6f>awoods: Are you going to reply to scossu?15:44
<awoods>amccarty: are versions involved?
ajs6f: yes, like I said.
<ajs6f>awoods: You did not say.
<awoods>ajs6f: yes
<dchandekstark>yes, we saw 'cannot locate child node within parent' but not in that exact context IIRC
<amccarty>awoods: and you've hit the limit of my knowledge... but I can find out, one sec.
<ajs6f>awoods: COnsider it said.
amccarty: Can you get the relevant party into IRC?15:45
* peichman joins
<dchandekstark>ajs6f escowles: fwiw, we saw that type of error in version creation - see https://jira.duraspace.org/browse/FCREPO-2076
although i'm not sure that was the only context in which that type of error occurred15:46
* bseeger joins
<amccarty>ajs6f: he's not around - but if there's an easy way to determine that, I can find out
<ajs6f>amccarty: Hm. I don't think there is without a running repo. Maybe you can give us alittle more context? How is this repo being used?15:47
<cm_harlow>ajs6f no versioning of resources there that i was aware of. was used for resources we were migrating from dlxs primarily. fed into a curation concerns hydra stack. don't know if this is what you're asking or helps tho.15:50
amccarty is the brains here so i apologize if i misunderstand a question. i just know about end user cases + what we were trying before this happened at a high level.15:51
<awoods>cm_harlow/amccarty: are you using LevelDB or a different backend database? Did a certain event take place after which there was corruption?15:54
<ajs6f>This definitely has the sulfurous stank of LevelDB.
<amccarty>awoods: leveldb, and not sure about an event
awoods: though I have heard about imports/migrations of collections crashing and having to be restarted15:55
so, there are those events
<awoods>amccarty: are you using any other JAVA_OPTS beyond what your servlet container comes with?
<ajs6f>amccarty: Crashing like the job didn't succeed, or crashing like the repo went down. Similarly, restarting like "Start the job again." or restariting like "Turn the server off and on."15:56
amccarty: Sorry, that was a question.
<amccarty>ajs6f: crashing like the job didn't succeed, and was restarted15:57
<ajs6f>amcarty: Okay, but the repo stayed ip?
<amccarty>not a server/tomcat crash
<ajs6f>up
amccarty: Okay, cool.
<amccarty>ajs6f: apparently it did, until someone tried restarting tomcat for some reason or another
<ajs6f>amccarty: Ah. Erm.15:58
<amccarty>awoods: other JAVA_OPTS: 'ADD_JAVA_OPTS="-XX:+UseG1GC -Xmx12288m -Xms256m -Dfcrepo.home=/cul/data/fedora-data -Xms1024m -XX:NewSize=512m -XX:MaxNewSize=512m -XX:PermSize=512m -XX:MaxPermSize=512m -XX:+DisableExplicitGC -Dfcrepo.log=DEBUG"'
<awoods>amccarty: what is your operating system?16:00
centos?
<amccarty>RHEL 6.7 x86_6416:01
<awoods>amccarty: It would be interesting if the scripts escowles referenced have a positive effect.16:03
<amccarty>awoods: let me run it now - it's got to process 25+GB on spinning disk, so might take awhile...;)
while that's running - it sounds like we should be using something other than leveldb, is that right?16:04
* github-ff joins16:05
[fcrepo-import-export] escowles force-pushed uri-translation from cbf22cf to 67e08a5: https://git.io/vi8ax
fcrepo-import-export/uri-translation 74273f2 Esmé Cowles: Translating URIs on import
fcrepo-import-export/uri-translation e109c9a Esmé Cowles: Addressing code review comments
fcrepo-import-export/uri-translation 2d848b6 Esmé Cowles: Flushing Triple Sink, adding Javadoc comment
* github-ff leaves
<awoods>amccarty: yes, MySQL or PostgreSQL16:07
<amccarty>awoods: any preference?
<awoods>amccarty: It depends on who you ask. escowles? mikeAtUVa?16:08
<ruebot>PostgreSQL has seen the best results in the performance and scaling tests so far.
<ajs6f>I think I can say that we would prefer PostgresQL, although I'm not sure we're actually using it now.
<ruebot>MySQL is fine as well, really depends on your experience with both of them, and how much configuration you want to do.16:09
<escowles>we're using postgres at princeton, but mostly because that's what we use generally
* dchandekstark leaves16:10
<ajs6f>awoods: The point about the PR is that URI.create is only a good thing when you don't want to deal with the meaningless exception because you are confident that it will not occur. We have no right to that confidence (we're getting RDF files off of disk) and we should catch the exception and deal with it, just as is being done.
<amccarty>ok cool - will start with mysql since it's standard here
* dchandekstark joins16:16
* manez leaves16:19
* dchandekstark leaves16:21
<amccarty>awoods: running that script in a screen - assuming if it can fix, it will do so16:27
* ajs6f leaves
<amccarty>I'm out for now, will be back later or tomorrow and will keep you guys posted16:31
* dchandekstark joins16:33
* amccarty leaves16:36
* dwilcox leaves16:40
* manez joins16:43
* manez leaves16:47
* dhlamb leaves16:48
* cm_harlow leaves16:51
* peichman leaves16:56
* bseeger leaves16:59
* apb18 leaves17:10
* youn joins17:14
* mikeAtUVa leaves17:21
* manez joins17:29
* manez leaves17:33
* acoburn leaves17:50
* whikloj leaves17:52
* dchandekstark leaves17:59
* apb18 joins18:29
* manez joins18:30
* manez leaves18:34
* apb18 leaves18:36
* apb18 joins20:19
* youn leaves20:43
* youn joins21:08
* apb18 leaves21:20
* mikeAtUVa joins21:33
* github-ff joins21:46
[fcrepo-import-export] awoods pushed 1 new commit to master: https://git.io/vi4tW
fcrepo-import-export/master e7dc46b Michael Durbin: Added authorization support for import/export. (#26)...
* github-ff leaves
* escowles leaves22:23
* escowles joins22:24
<youn>to test indirect containers, would it make sense to create a collection for publications with proxies for publications that are currently contained in their authors (by virtue of how lubm identifiers are set up)?22:38
* mikeAtUVa leaves22:45
* youn leaves23:28

Generated by Sualtam