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

Using timezone: Eastern Standard Time
* dbernstein leaves01:37
* dbernstein joins01:38
* dbernstein leaves02:20
* dbernstein joins02:21
* dbernstein leaves02:22
* dwilcox joins02:47
* thomz joins05:01
* dwilcox leaves05:48
* dwilcox_ joins
* dwilcox_ leaves06:31
* acoburn joins08:24
* dwilcox joins08:37
* whikloj joins08:54
* dwilcox leaves09:07
* dwilcox_ joins
* mikeAtUVa joins09:08
* dwilcox_ leaves09:18
* ajs6f joins09:28
* osmandin joins09:35
* mohamedar joins09:47
* dwilcox joins09:57
* dwilcox leaves10:01
* setsquare joins10:03
* youn joins10:07
<ajs6f>Just made a tech meeting page:10:12
Feel free to add stuff.
* dhlamb joins10:14
* osmandin leaves10:17
* dbernstein joins10:22
* setsquare leaves10:40
* github-ff joins10:44
[fcrepo4] acoburn opened pull request #1111: Make it possible to set the location for velocity.log (master...fcrepo-2260) https://git.io/vPC8E
* github-ff leaves
* dwilcox joins10:54
<ajs6f>Time for https://wiki.duraspace.org/display/FF/2016-10-06+-+Fedora+Tech+Meeting10:58
* ajs6f is here
* ruebot is here
* whikloj is here11:00
* dwilcox leaves11:01
* mikeAtUVa is here, but will have to depart at 11:40.
<acoburn>*is here*
* dhlamb is here
* dwilcox joins11:02
<ruebot>temple iirc11:03
<whikloj>University of Pennsylvania Libraries - Katherine Lynch - katherly@upenn.edu11:04
<dhlamb>whikloj: totally not creepy
<whikloj>dhlamb: I excel at not creepy
* youn leaves11:05
* dshalvi joins
<ruebot>but, it is in HD!
* dwilcox_ joins11:06
* dwilcox leaves
* dsteelma joins11:08
* dsteelma leaves11:10
* thomz leaves
* dwilcox_ leaves11:11
'all kind of blurry and really colorful'
<ajwagner>whikloj: I can most likely take time to run some tests on it tomorrow. (But then I'm on vacation for 2+ weeks)
<ajs6f>Fedora 4: YMMV11:12
<ajwagner>I will try to be an experimentalist, to see if all this high-falutin' theory works out ;)11:13
<ajs6f>mikeAtUVa: Can we do that?
* dwilcox joins
I'm just using those for the MODESHAPE 4->5 test
* dwilcox_ joins11:18
* dwilcox leaves11:20
<ajs6f>That's because acoburn is optimistic about changing the world for the better.
<dhlamb>i think i followed that11:21
<ajwagner>acoburn++ #this is great work
* dwilcox_ leaves11:23
* dwilcox joins
acoburn: jump in, sounds like you've got it11:35
<dhlamb>give 'er
<whikloj>dbernstein: embrace the flood11:36
<ajs6f>whikloj: Is that from the old Jung Fu TV show?
Kung Fu.
<whikloj>ajs6f: Be reed in the wind, grasshopper11:37
<ajs6f>Although I would like to see Jung Fu. That sounds like a great show involving people battling their own archetypes.
<whikloj>Much nicer than anything ajs6f has said about it11:41
<dhlamb>software development is incredibly human, to quote acoburn11:42
java without spring11:43
that's the wild west
<ruebot>...there is a LDP Next slack if folks are interested
<ajwagner>^ link Nick?
We've (I've) been thinking a lot about LDN.11:46
(at U of T)
<ruebot>email cody.burleson@base22.com11:47
<ajs6f>https://kafka.apache.org/ not https://en.wikipedia.org/wiki/Franz_Kafka11:49
* dwilcox leaves
* dwilcox joins11:51
* dwilcox leaves11:55
<whikloj>PCDM call in 3 minutes11:57
<ajs6f>whikloj: thnx11:58
* whikloj is dropping # Cheers all12:00
<dhlamb>gotta go to PCDM call
I'll check out the PR acoburn
* escowles joins12:11
* escowles leaves
<acoburn>ajs6f: can you check if I got the participant list right? https://wiki.duraspace.org/display/FF/2016-10-06+-+Fedora+Tech+Meeting12:18
<ajs6f>acoburn: Yeah, that seems right to me.12:19
* escowles joins12:25
* escowles leaves12:59
<acoburn>ajs6f: I have a log4j question if you don't mind13:11
ajs6f: I've noticed that the jbossjta (transaction) jar has a log4j configuration baked in, which writes a transaction.log to the current directory (sort of like the velocity.log issue)13:12
ajs6f: I've found a work-around, but it seems flaky and I'd like some advice13:13
* escowles joins
* escowles leaves13:17
* ajs6f leaves13:53
* escowles joins14:09
* ajs6f joins14:26
* ajs6f leaves
* ajs6f joins
acoburn: You can't just exclude log4j from that dep and make sure we have log4j-over-slf4j (or whatever it is called) online?14:27
<acoburn>ajs6f: I figured it out — my tomcat had log4j jars in the $CATALINA_HOME/lib directory14:28
<ajs6f>acoburn: Ah… CLASSPATHS RULZ
<acoburn>ajs6f: so I could either remove those jars or override the log4j configuration via system properties14:29
<ajs6f>acoburn: Yeah. What's the solution for the public?
<acoburn>ajs6f: I think is sorta depends — installing tomcat via yum on centos doesn't include log4j jars by default14:30
ajs6f: but some installations may have a need for that
<ajs6f>mikeAtUVa: Did you see Kevin Ford's message to the list?
mikeAtUVa: Is that the bug you fixed?
acoburn: Document and punt?14:31
<acoburn>ajs6f: yes
ajs6f: re Kevin's message, yes, that's the bug mikeAtUVa fixed
ajs6f: there are some merge conflicts with the PR though
<ajs6f>acoburn: Meaning we haven't merged it?14:32
<acoburn>ajs6f: I think we discussed not putting that into 4.7.0 b/c it's a potentially big change, but I think it can go into the current master whenever
ajs6f: yes, it's not merged
ajs6f: seems like nothing gets merged when awoods is out
<ajs6f>mikeAtUVa: Can you respond to Kevin on-list with an explanation?14:33
* osmandin joins14:44
* mjgiarlo joins14:47
* osmandin leaves14:51
<mikeAtUVa>ajs6f: yeah, I'll respond... I'm (figuratively) puttin' out a fire now, but I'll do it next.14:56
<ajs6f>mikeAtUVa++ # hope it's not too big a fire14:57
Fires are annoying. mikeAtUVa and I just had to go outside for a fire alarm.
* escowles leaves14:58
* ruebot leaves15:08
* dwilcox joins15:46
* dshalvi leaves15:47
* kefo joins15:49
<ajs6f>Hello, kefo!
<kefo>I've got a question about upgrading from fcrepo 4.6 to fcrepo 4.7, which includes a shiny new version of modeshape.
Is there something special I need to do to migrate an existing repository to this new version?15:50
I installed fcrepo4.7rc1 and can get it working, but fcrepo4.7rc1 creates a new table in Postgres. That new table doesn't match the old table in design. So I feel like I missed a step.15:51
<ajs6f>kefo: Yes. Yes, there is.
kefo: It is a _complete_ migration. The backends are not compatible.
kefo: You will have to export/reimport.
<acoburn>kefo: you'll want to run the fcr:backup on the repo
<ajs6f>kefo: Sorry.15:52
<acoburn>kefo: and then run fcr:restore
<acoburn>kefo: there is also one other step related to configuration. I'll dig that up in a sec...15:53
<kefo>Is this documented yet. :)
The need, that is.15:54
<acoburn>kefo: you probably have a system property for fcrepo.modeshape.configuration
* dhlamb leaves
* mjgiarlo leaves
<acoburn>kefo: you'll need to make sure that it points to the right place for 4.7.0
<whikloj>acoburn: This? https://wiki.duraspace.org/display/FEDORA4x/Application+Configuration15:55
<acoburn>whikloj: yes
kefo: there was a built-in config for 4.6.0 (minimal-default?) that is not part of 4.7.0 — I think that may have used leveldb15:56
kefo: but if you're not already using a database (postgresql/mysql) for the backend, this would be a good time to move to that15:57
<kefo>I dug up the new/changed repository.json file (where fcrepo.modeshape.configuration points) and made the relevant changes.
<ajs6f>kefo: What acoburn says, big time. Get to a RDB.
<acoburn>kefo the process would be something like: run fcr:backup on the existing repo15:58
kefo: then shut down the servlet container
<kefo>Oh, we're on postgres. Have been for a while.
<kefo>Right. I'm just considering a full data reload.
<ajs6f>kefo: Yeah, fun.
<kefo>Versus backup/restore.
<acoburn>kefo: if a full data reload is an option, it may be a better choice
kefo: I've only used backup/restore on relatively small repos
kefo: and I've heard that for large-ish repos it can take a _really_ long time16:00
<kefo>Hmm… that's helpful @acoburn16:01
<acoburn>kefo: I'm not sure how the two approaches compare speed-wise, but if it were me, I'd do the data re-load
<kefo>I'm leaning that way. It's just cleaner and also already to go.16:02
<ajs6f>The JCR-based backup/restore means a _massive_ amount of XML parsing.
<acoburn>kefo: the reason for that is just that I don't trust modeshape-on-infinispan (on a single node, that is), and for a _really_ long-running job, I worry that something would go wrong
<kefo>And we're talking about something on the order of 200,000 resources.
<acoburn>kefo: ok, that's not too bad size-wise, but it would also be relatively easy to re-ingest them; the main issue to consider is how the URIs are generated16:04
* mohamedar leaves16:05
<acoburn>kefo: if you need the resource URLs to be consistent from 4.6 to 4.7 and you've been relying on POST commands (which generate a UUID-like identifier), you would get different values upon re-ingesting
kefo: that may or may not be an issue, depending on how those URIs are exposed to the public internet (or not)16:06
<kefo>We mint the URIs and they're predictable each time. Resources go in using PUTs.16:07
<acoburn>kefo: then you've clearly designed the system well
<whikloj>ajs6f/acoburn: This raises a step I better move on, how does one cut a new branch of the documentation?16:09
<kefo>I'd have done the same (I like to control my URIs), but the credit in this case goes to Stefano.
<acoburn>whikloj: if it's a Confluence question, the answer is ask awoods (I have no idea)
<ajs6f>whikloj: It's a copy, much like making a new page for a tech call. But I don't know that you will have permissions for that.
<whikloj>acoburn++ # It is and I will thanks16:10
<ajs6f>whikloj: Yeah, ping awoods.
<whikloj>ajs6f: I do, but do you copy all the pages individually?
<ajs6f>whikloj: No, it's a new space. I doubt you have the permissions. Can you make new spaces?16:11
<acoburn>kefo: oh, are you Kevin? I didn't think about the handle
<kefo>It is I.
<ajs6f>He's either Kevin Ford, or an incredible simulation.
<whikloj>ajs6f: no I can't... awoods it is16:12
* ajs6f leaves16:20
So, despite your good advice, we are looking at the backup/restore method.16:32
Mostly because we'd like to know how long it takes, what is required, etc. We have the luxury to test this.16:33
I'm looking at good practices, such as backing up the database and the fcrepo data directories.
I find myself with a few questions on this last point.
1) Will the size of fcrepo:backup be, more or less, equal to the fcrepo data directories?16:34
2) What is the role the 'trash' directory, for example, in the frepo.binary directory?16:35
3) I loaded a goodly amount of images in to the repository. I then deleted them all, by issuing a DELETE on the first level of pairtrees under the primary container. I deleted the tombstones also. I get a 404 for all of these. How come the fcrepo.binary directory still contains these files?16:39
<acoburn>kefo: when you go to restore the data, your database schema will be completely different, so it's probably best to use a separate pgsql schema/database
kefo: re. the binary directory containing deleted images, that has to do with how modeshape handles deleted resources16:40
<kefo>Oh, I know. I'm grabbing the backup in case, for some reason, something awful happens and we have to revert to 4.6
<acoburn>kefo: once a day (that's the default) a thread finds all the delete binaries and _actually_ deletes them16:41
<kefo>So, if I wanted to backup the directory itself, it would be wiser to wait 24 hours, yes?
<acoburn>kefo: maybe? it depends on how much data may be involved16:42
<kefo>If I performed a fcr:backup, deleted resources, whether modeshape is still tracking them or not, would not be backed up. Yes/no?
Roughly 856 Gb.
<acoburn>kefo: my understanding is that those deleted resources would _not_ be in the backup
kefo: if they are, there is something wrong with modeshape16:43
<kefo>OK. Thanks. I think I can make a plan from this.16:47
* mikeAtUVa leaves17:02
* acoburn leaves17:21
* dwilcox leaves17:26
* whikloj leaves17:59
* kefo leaves19:07
* mjgiarlo joins20:31
* mjgiarlo leaves20:32
* mjgiarlo joins
* mjgiarlo leaves20:33
* mjgiarlo joins20:34
* mjgiarlo leaves20:35
* mjgiarlo joins
* mjgiarlo leaves20:36
* mjgiarlo joins20:37
* mjgiarlo leaves
* mjgiarlo joins20:38
* mjgiarlo leaves20:39
* mjgiarlo joins20:40
* mjgiarlo leaves
* mjgiarlo joins20:41
* mjgiarlo leaves20:42
* mjgiarlo joins20:43
* mjgiarlo leaves
* mjgiarlo joins20:44
* mjgiarlo leaves20:45
* mjgiarlo joins
* mjgiarlo leaves20:46
* mjgiarlo joins
* mjgiarlo leaves20:47
* mjgiarlo joins20:48
* mjgiarlo leaves
* mjgiarlo joins20:49
* mjgiarlo leaves20:50
* mjgiarlo joins
* mjgiarlo leaves20:51
* mjgiarlo joins
* mjgiarlo leaves20:52
* mjgiarlo joins
* mjgiarlo leaves20:53
* mjgiarlo joins20:54
* mjgiarlo leaves20:55
* mjgiarlo joins20:58
* mjgiarlo leaves
* mjgiarlo joins20:59
* mjgiarlo leaves21:00
* mjgiarlo joins
* mjgiarlo leaves21:01
* mjgiarlo joins21:02
* mjgiarlo leaves21:03
* mjgiarlo joins
* mjgiarlo leaves21:04
* mjgiarlo joins
* mjgiarlo leaves21:05
* mjgiarlo joins21:06
* mjgiarlo leaves
* mjgiarlo joins21:07
* mjgiarlo leaves21:08
* mjgiarlo joins21:09
* mjgiarlo leaves
* mjgiarlo joins21:10
* mjgiarlo leaves21:11
* mjgiarlo1 joins21:14
* mjgiarlo1 leaves21:15
* mjgiarlo1 joins21:16
* mjgiarlo1 leaves
* mjgiarlo joins21:17
* mjgiarlo leaves21:18
* mjgiarlo joins
* mjgiarlo leaves21:20
* mjgiarlo joins21:21
* mjgiarlo leaves21:22
* mjgiarlo joins21:23
* mjgiarlo leaves
* mjgiarlo joins21:25
* mjgiarlo leaves
* mjgiarlo joins21:26
* mjgiarlo leaves21:27
* mjgiarlo joins21:41
* mjgiarlo leaves21:42
* mjgiarlo joins
* mjgiarlo leaves21:43
* mjgiarlo joins21:44
* mjgiarlo leaves21:45
* mjgiarlo joins
* mjgiarlo leaves21:46
* mjgiarlo joins21:47
* mjgiarlo leaves
* mjgiarlo joins21:48
* mjgiarlo leaves
* mjgiarlo joins21:49
* mjgiarlo leaves21:50
* mjgiarlo joins
* mjgiarlo leaves21:51
* mjgiarlo joins21:52
* mjgiarlo leaves21:53
* mjgiarlo joins
* mjgiarlo leaves21:54
* mjgiarlo joins21:55
* mjgiarlo leaves21:56
* mjgiarlo joins21:57
* mjgiarlo leaves
* mjgiarlo joins21:58
* mjgiarlo leaves21:59
* mjgiarlo joins
* mjgiarlo leaves22:00
* mjgiarlo joins22:01
* mjgiarlo leaves22:02
* mjgiarlo joins22:03
* mjgiarlo leaves
* mjgiarlo joins22:04
* mjgiarlo leaves
* mjgiarlo joins22:05
* mjgiarlo leaves
* mjgiarlo joins22:06
* mjgiarlo leaves22:07
* mjgiarlo joins22:08
* mjgiarlo leaves22:09
* mjgiarlo joins22:10
* mjgiarlo leaves22:11
* mjgiarlo joins22:12
* mjgiarlo leaves
* mjgiarlo joins22:13
* mjgiarlo leaves22:14
* mjgiarlo joins
* mjgiarlo leaves22:15
* mjgiarlo joins22:16
* mjgiarlo leaves22:17
* mjgiarlo joins22:18
* mjgiarlo leaves
* mjgiarlo joins22:19
* mjgiarlo leaves22:20
* mjgiarlo joins
* mjgiarlo leaves22:21
* mjgiarlo joins22:22
* mjgiarlo leaves22:23
* mjgiarlo joins22:24
* mjgiarlo leaves
* mjgiarlo joins22:26
* mjgiarlo leaves
* mjgiarlo joins22:27
* mjgiarlo leaves
* mjgiarlo joins22:29
* mjgiarlo leaves
* mjgiarlo joins22:30
* mjgiarlo leaves22:31
* mjgiarlo joins
* mjgiarlo leaves22:32
* mjgiarlo joins22:33
* mjgiarlo leaves22:34
* mjgiarlo joins22:35
* mjgiarlo leaves
* mjgiarlo joins22:36
* mjgiarlo leaves
* mjgiarlo joins22:38
* mjgiarlo leaves
* mjgiarlo joins22:39
* mjgiarlo leaves
* mjgiarlo joins22:40
* mjgiarlo leaves22:41
* mjgiarlo joins22:42
* mjgiarlo leaves
* mjgiarlo joins22:43
* mjgiarlo leaves
* mjgiarlo joins22:45
* mjgiarlo leaves
* mjgiarlo joins22:46
* mjgiarlo leaves22:47
* mjgiarlo joins
* mjgiarlo leaves22:48
* mjgiarlo joins22:49
* mjgiarlo leaves
* mjgiarlo joins22:51
* mjgiarlo leaves
* mjgiarlo joins22:52
* mjgiarlo leaves
* mjgiarlo joins22:53
* mjgiarlo leaves22:54
* mjgiarlo joins
* mjgiarlo leaves22:55
* mjgiarlo joins22:56
* mjgiarlo leaves22:57
* mjgiarlo joins
* mjgiarlo leaves22:58
* mjgiarlo joins
* mjgiarlo leaves22:59
* mjgiarlo joins23:00
* mjgiarlo leaves
* mjgiarlo joins23:01
* mjgiarlo leaves23:02
* mjgiarlo joins
* mjgiarlo leaves23:03
* mjgiarlo joins23:04
* mjgiarlo leaves
* mjgiarlo joins23:05
* mjgiarlo leaves23:06
* mjgiarlo joins23:07
* mjgiarlo leaves23:08
* mjgiarlo joins23:09
* mjgiarlo leaves23:10
* mjgiarlo joins
* mjgiarlo leaves23:11
* mjgiarlo joins23:12
* mjgiarlo leaves23:13
* mjgiarlo joins23:14
* mjgiarlo leaves

Generated by Sualtam