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

Using timezone: Eastern Standard Time
* dbernstein leaves02:15
* thomz joins03:34
* escowles_ leaves06:50
* escowles joins
* dwilcox joins08:23
* whikloj joins09:06
* ajs6f joins09:19
* peichman joins09:25
* mikeAtUVa joins09:26
* acoburn joins09:34
<whikloj>barmintor: Would it be helpful if the vagrant setup had a MySQL and/or Postgres option?
<ajs6f>barmintor: ping09:36
whikloj: What does the Vagrant thing do now?
<whikloj>ajs6f: it uses the file-simple backing, which I don't understand
<ajs6f>whikloj: I never heard of that, but I suppose it just stores stuff in some kind of journals?09:37
<whikloj>ajs6f: Yeah not sure, not for production I'm sure
ajs6f: https://github.com/fcrepo4/fcrepo4/blob/master/fcrepo-configs/src/main/resources/config/file-simple/repository.json09:38
<ajs6f>whikloj: Well, the Vagrant box isn't for produciton, right?
<whikloj>ajs6f: right, but then barmintor had asked about vagrant and I starting thinking about testing. Would it help if standing up a MySQL configured Fedora setup was easier09:40
<ajs6f>whikloj: Hm. Depends on what we are testing and to what extent testing against a Vagrant box is testing Vagrant instead of what we want to test.
whikloj: I don't know much about Vagrant, so maybe that's not a real concern.09:41
<whikloj>ajs6f: yeah vagrant is just a virtualized machine, in our case an Ubuntu linux box.
* th5 joins09:42
<whikloj>ajs6f: or rather vagrant is a way to provision a virtualized machine, it actually runs on VirtualBox I think
<ajs6f>whikloj: Sure, but a given Vagrant setup also includes a bunch of scripts that install stuff right? So then we are testing that too. (You can use different "providers" of virtualization with Vagrant, but VB is the most popular by far, as I understand it.)
https://www.vagrantup.com/docs/providers/index.html09:43
"While Vagrant ships out of the box with support for https://www.virtualbox.org/, https://www.microsoft.com/hyper-v, and https://www.docker.io/, Vagrant has the ability to manage other types of machines as well. This is done by using other providers with Vagrant."
So do we end up testing _that_, too?
<whikloj>ajs6f: no, I see vagrant as an environment to test Fedora. I'm not looking to support all the virtualization machinery. I just thought it might be beneficial to have a way to standup a Fedora using MySQL/Postgres.09:45
<awoods>whikloj: I think you raise a good point about having mysql on the vagrant... especially since it is easy to script the install.
whikloj: we currently use file-simple because it is, well, simple09:46
whikloj: is there a downside to switching to mysql or postgres?
<whikloj>awoods: It does add a level of complexity in repository.json. If you want MySQL and WebAC, we need to provide a repository.json because I don't think there is one we can point at.09:47
<awoods>whikloj: yes, I do that locally
<whikloj>awoods: otherwise, I'm not sure. I don't know that there is any particular upside or downside.09:48
<ajs6f>whikloj:awoods: I'm in no way opposed to switching to something more "real" for the backend on Vagrant. I'm saying that using Vagrant "as an environment to test Fedora" as whikloj suggests necessarily means we would be testing all the things that could be variable or just wrong with a Vagrant setup in addition to what we are trying to test. whikloj, can you articulate the gain a little more clearly?
<awoods>ajs6f: I agree, just because we have mysql or postgres or something else built into vagrant, does not mean it is the canonical testing platform by any means.09:49
ajs6f: like you say, it is just more real
<ajs6f>keeping_it_real++09:50
<whikloj>ajs6f/awoods: Yeah, a pass in vagrant should never be considered the "end all, be all" of quality control
ajs6f/awoods: But no one should be using Fedora with file-simple, so this would more closely mimic a "real" install
<ajs6f>whikloj: +1 to using on Vagrant what we recommend people use in life.09:51
* osmandin joins
<ajs6f>I use Apache Derby in-memory to test Fedora 3 components, because it is fast. I wonder if that's worth exploring for workshop setups. Can you pass flags of some kind into Vagrant, or does that get unworkably complex?09:52
<whikloj>ajs6f: It will read environment variables, thought we general use a config file.
https://github.com/fcrepo4-exts/fcrepo4-vagrant/blob/master/install_scripts/config09:53
<ajs6f>whikloj: I dunno. It all comes down to "what is the vagrant box for?" which I don't really know.09:55
* osmandin leaves10:07
* ajs6f leaves10:08
<barmintor>someone pinged me, I have 10 minutes: what's up?10:09
* thomz leaves10:12
* youn joins10:15
* ajs6f joins10:18
* bseeger joins10:26
* kefo joins10:27
* github-ff joins10:48
[fcrepo4] acoburn closed pull request #1131: Fcrepo 2248: Adds missing Content-Length header on HEAD request for non binary obj (master...fcrepo-2248) https://git.io/vXk98
* github-ff leaves
* ajs6f leaves11:00
* ajs6f joins11:04
<whikloj>Okay I always have trouble with this. In the fcrepo4-config for mysql, it uses properties like "username" : "${fcrepo.mysql.username}",11:14
But in the wiki page it says : JAVA_OPTS="${JAVA_OPTS} -Dfcrepo.ispn.mysql.username=<username>"
* bseeger leaves
<acoburn>whikloj: isn't that due to the 4.6.0 -> 4.7.0 changes11:15
<whikloj>acoburn: Ummm I'm not sure, is it? Was the ispn removed in 4.7.0?11:16
<acoburn>whikloj: yes, that was part of the MODE4 -> MODE5 change
whikloj: of course, it's still a good idea to _not_ use JAVA_OPTS for that data11:17
<whikloj>acoburn: ok, thank you. I've had this problem before but I'm never sure what I might be doing wrong.
<acoburn>whikloj: anyone with access to the server can run `ps ax` to get those values
<whikloj>acoburn: Yeah I'm using CATALINA_OPTS for vagrant, but for production I'm guessing that JNDI is recommended
<acoburn>whikloj: as long as those values are managed in such a way that they're not visible to everyone w/ access to the server (that is, if you care about such things)11:18
<whikloj>acoburn: So a question about the Performance Note on this page :https://wiki.duraspace.org/display/FEDORA47/Configuring+JDBC+Object+Store11:19
acoburn: Is that still relevant?
<acoburn>whikloj: it's accurate, I just wouldn't recommend it11:20
<whikloj>acoburn: ok thanks, just wanted to make sure it still applied
<awoods>whikloj: are you testing kefo's issue?11:21
<whikloj>awoods: I'm just trying to setup a MySQL/PostgresSQL vagrant, then perhaps I can try.
<awoods>whikloj: thanks. I am in that same space as well... at least testing 4.7 / 4.6-patch with mysql and kefo's data.11:22
* ajs6f leaves11:42
* ajs6f joins11:44
* ajs6f leaves11:45
* jjtuttle leaves11:48
* jjtuttle joins11:49
<whikloj>If I can connect to MySQL but Fedora4 can't using the same credentials, port, etc...suggestions on where to start?11:57
* dbernstein joins12:08
<awoods>whikloj: what error are you getting from F4?12:10
<whikloj>awoods: nested exception is org.fcrepo.kernel.api.exception.RepositoryRuntimeException: com.zaxxer.hikari.pool.HikariPool$PoolInitializationException: Exception during pool initialization: HikariPool-0 - Connection is not available, request timed out after 30000ms.12:11
awoods: further down on the second stack trace I see12:12
Caused by: com.mysql.jdbc.exceptions.jdbc4.MySQLSyntaxErrorException: Access denied for user 'fedora_user'@'localhost' to database 'fcrepo'
<awoods>whikloj: and you are sure your credentials are correct?12:13
* dwilcox leaves
<whikloj>awoods: I can connect using `mysql -hlocalhost -P3306 -ufedora_user -pfedora_passwd fedora`
awoods: I am sure I am missing something obvious12:14
OMG
<awoods>whikloj: I get that error when my password is incorrect12:16
<whikloj>awoods: you are the finder of all my facepalms
awoods: I created the database "fedora" in the script not "fcrepo"12:17
I just couldn't see it
* bseeger joins12:19
<awoods>whikloj: good to hear you are moving forward12:27
* bseeger leaves12:28
* peichman leaves13:00
* peichman joins13:01
* github-ff joins13:20
[fcrepo4-vagrant] whikloj opened pull request #69: Allow for a MySQL backend (master...mysql-backend) https://git.io/vXlLV
* github-ff leaves
* dwilcox joins13:47
* dbernstein leaves13:49
* dbernstein joins13:50
* peichman leaves13:58
* peichman joins13:59
* osmandin joins14:27
* ajs6f joins14:31
* youn leaves14:46
* osmandin leaves15:23
<whikloj>kefo: What version of python are you running for your scripts?15:42
<kefo>3.515:43
<whikloj>kefo: thanks, 2.7 had a hissy fit so I was guessing it was 3.x15:48
kefo: So the load_resources ingests a bunch of objects and then GETs them all to verify?15:58
<kefo>whikloj: load_resources loads the resources. test-gets performs the GETs for verification purposes.15:59
whikloj: You'll need to set the —data param for test-gets. https://github.com/aic-collections/fcrepo-backup-restore-tests#running16:00
<whikloj>kefo: ok, it added a bunch and now it is sitting there... possibly waiting for the triplestore indexer to catch up16:01
* acoburn1 joins16:02
* acoburn leaves16:03
<kefo>whikloj: Dunno. Once it PUTs the resources, it ends by spitting out some details about the load overall. Ditto for the GETs; the end is marked by general info about the run.16:04
<whikloj>kefo: ok, I'll wait. It is a small test machine and I still see action on the Fuseki side so something is happening.16:05
* dwilcox leaves
<kefo>whikloj: This is probably immaterial, or certainly should be, but I've not had anything else hooked up. Just a standalone fedora.
* dwilcox joins16:06
<whikloj>kefo: I'll remember to turn that off next time :)
* dwilcox leaves16:10
* dbernstein_ joins16:42
* dbernstein leaves16:48
* th5 leaves17:04
* acoburn1 leaves17:08
* ajs6f leaves17:18
* peichman leaves17:30
* whikloj leaves18:00
* dbernstein leaves18:54
* dbernstein joins18:55
* kefo leaves
* dwilcox joins18:58
* dwilcox leaves19:06
* dbernstein leaves20:30
* mikeAtUVa leaves20:44
* github-ff joins20:53
[fcrepo4-upgrade-utils] awoods created revert-14-master (+1 new commit): https://git.io/vX8Tl
fcrepo4-upgrade-utils/revert-14-master 39f71df Andrew Woods: Revert "Added prerequisites; Reorganized and finessed BackupFixer description…"
* github-ff leaves
* mikeAtUVa joins20:59
* travis-ci joins21:09
fcrepo4-exts/fcrepo4-upgrade-utils#13 (revert-14-master - 39f71df : Andrew Woods): The build passed.
Change view : https://github.com/fcrepo4-exts/fcrepo4-upgrade-utils/commit/39f71dfce3d1
Build details : https://travis-ci.org/fcrepo4-exts/fcrepo4-upgrade-utils/builds/173425378
* travis-ci leaves
* f4jenkins joins23:12