Time
S
Nick
Message
01:44 jri joined #dataverse
02:25 yoh joined #dataverse
04:00 jri joined #dataverse
07:11 jri joined #dataverse
07:25 jri joined #dataverse
11:12 donsizemore joined #dataverse
11:43 donsizemore
@pdurbin I think https://github.com/IQSS/dataverse-ansible/issues/22 is good to go. want to test it before i merge? for vagrant, you'd add a branch name to tests/group_vars/vagrant.yaml
11:52 pdurbin
donsizemore: great! I won't be able to test it for a couple hours so if you're itching to merge it, please go ahead.
13:13 donsizemore joined #dataverse
13:27 * pdurbin
edits the YAML and runs `vagrant up`
13:39 pdurbin
donsizemore: "build warfile" is taking a while
13:40 pdurbin
tail'ing /tmp/dataverse/mvn.out assures me that things are happening :)
13:41 donsizemore
@pdurbin yis. i had been opening a separate terminal and tail -f -ing /tmp/dataverse/mvn.out . i can possibly make ansible spit out the contents of that file if you'd like me to
13:41 pdurbin
That's ok. I did the same. New terminal. `vagrant ssh`. `tail`.
13:43 pdurbin
BUILD SUCCESS
13:43 donsizemore
note that i told maven to skip tests during the build, dunno if you care about that.
13:44 pdurbin
That's fine. Did you see we added Maven profiles to control which tests are run in https://github.com/IQSS/dataverse/pull/4989 ?
13:48 pdurbin
Oh. Shoot. I already had something on port 8080. Glassfish running on my Mac.
13:52 pdurbin
donsizemore: can I just do a `vagrant halt` and change the port and then do `vagrant up` again? I guess I'll try that.
13:52 julian joined #dataverse
13:53 donsizemore
@pdurbin you can probably get away with that. i always do a destroy and start over. 20 year-old me would compile bleeding-edge linux kernels for my workstation. 40 year-old me just wants stuff to work.
13:58 * pdurbin
crosses fingers
14:00 pdurbin
Hey, it worked.
14:00 pameyer joined #dataverse
14:43 jri joined #dataverse
14:57 pdurbin
I left a comment about it working at https://github.com/IQSS/dataverse/issues/4990#issuecomment-415425816
14:57 pdurbin
pameyer: thanks for installing AWS CLI . Any thoughts on the other two issues?
14:57 pdurbin
- "Why 2 buckets?"
14:58 pdurbin
- different failure notification style
14:58 pameyer
"why 2 buckets?" - I don't know why there are 2 buckets
14:59 pameyer
"different notification style" - hadn't notice it before; if it's consistent that notifications reporting a problem have a dataverse icon it may not be something that needs fixing
15:01 pdurbin
Yeah, I dunno. I guess I'll find out and standup soon about the style thing.
15:03 pdurbin
Do you think the two buckets are dataverse.files.s3-bucket-name vs dataverse.files.dcm-s3-bucket-name?
15:06 pameyer
I suspect that something with the import API needed to move things from one bucket to the internal dataverse bucket; and that the dcm is putting things into the first one prior to the api call
15:07 pdurbin
ok
15:07 pdurbin
thanks
16:21 pdurbin
donsizemore pameyer: what's your timeline on moving from EZID to DataCite?
16:23 pameyer
pdurbin: shortly after I get credentials
16:27 pdurbin
Ok, the InReviewWorkflowIT test failed on phoenix with "Dataset cannot be edited due to dataset lock". Couldn't reproduce it using docker-aio (uses EZID). Commit 438c6e6. Able to reproduce it on my laptop (like phoenix, uses DataCite). Redeployed the war file and now I can't reproduce it on my laptop. Rerunning the tests on phoenix.
16:27 pdurbin
¯\_(ツ)_/¯
16:30 jri joined #dataverse
16:33 pameyer
pdurbin: that's develop?
16:33 pameyer
weird
16:34 pdurbin
passed on phoenix this time
16:34 pdurbin
I guess I'm chasing ghosts.
16:39 pameyer
strangeness happenes sometimes
16:49 pdurbin
Yeah. At standup I said I'd create an issue but I guess I won't now. But I wanted to tell someone what I did so thanks for listening. :)
16:50 pameyer
sporadic errors :<
16:50 pameyer
on the one hand, they fix themselves. but on the other hand, they'll break again whenever they feel so inclined
16:53 pdurbin
heisenbugs
16:54 pdurbin
pameyer: meanwhile, I messed around with your pull request about braces: https://github.com/IQSS/dataverse/pull/4995
16:56 pameyer
pdurbin: I saw, thanks
16:56 pameyer
sorry about the typo
17:12 pameyer
and next time I look into a code formatter, I should check to see if it supports partial reformatting :(
17:14 pameyer
still need to sort that for https://github.com/IQSS/dataverse/pull/4991
17:21 pdurbin
I've never heard of a such a thing but it sounds nice.
17:29 donsizemore
@pdurbin Odum's move to DataCite will be up to Jon and Thu-Mai, and they're both out this week
17:31 pdurbin
Ok. And I'm out all next week. And a good part of the week after. So you guys are in charge. :)
17:51 donsizemore
Berlin? Australia? Botswana?
17:52 pdurbin
hmm? :)
17:52 donsizemore
(those are the upcoming conferences i know of)
17:54 pdurbin
Oh. Staycation.
17:56 pdurbin
pameyer: still there?
17:59 pameyer
pdurbin: just back
17:59 pdurbin
pameyer: I have your closeQuietly pull request up and I'm happy to reformat the code if you want. I did a couple classes already.
18:01 pameyer
pdurbin: if you're already in there, that'd be great. it's more or less hit it's time box for me today
18:01 pameyer
if not, I might get a chance tomorrow
18:02 pdurbin
Let me see what I can do. Are all of these changes necessary? BatchServiceBean?!?
18:04 pameyer
not 100% sure - I was closing things as I went along, and stopped when each page load didn't leak another 150 descriptors
18:04 pameyer
I didn't backtrack; but I'm relatively sure they're all potential leaks
18:05 pameyer
also relatively sure there are more lurking
18:06 pdurbin
Ok. How were you exercising BatchServiceBean? By importing datasets?
18:08 pameyer
I wasn't - that one was a static analysis one
18:08 pdurbin
Ok. What tool are you using?
18:09 pameyer
I was measuring dataset.xhtml, and using sonar and infer to try to narrow down possible leaks
18:09 pdurbin
Never heard of infer.
18:09 pameyer
neither found the exporter leak - that was from following with jdb
18:10 pameyer
https://github.com/facebook/infer
18:11 pdurbin
Thanks. As a rule I don't like passing changes to QA that I haven't tested myself. If we aren't exercising the code (BatchServiceBean), maybe we should back those changes out (but leave a TODO).
18:11 pdurbin
You may recall the "lots of ugly code" pull request from last summer. It drove me crazy.
18:12 pameyer
sounds reasonable to me
18:12 pameyer
my thinking was to get it so that loading a dataset page wasn't a cause of instability
18:12 pdurbin
Yes. Absolutely. That's a case of you exercising the code.
18:13 pdurbin
You probably are loading CustomizationFilesServlet by hitting the dataset page.
18:14 pameyer
didn't trace that one with the debugger
18:18 donsizemore
@pdurbin your reference to last summer's PR has me giggling
18:19 pdurbin
It's true. The code was ugly. Still is. But let's not change it without knowing how to even exercise it.
18:20 pameyer
@donsizemore I think that was the actual title of the PR
18:20 donsizemore
@pameyer it was =)
18:29 pdurbin
pameyer: man, I'm losing steam on this thing.
18:33 pameyer
pdurbin: fair enough. thanks for giving it a shot
18:34 pdurbin
I'm not done. Let me push what I've got soonish.
18:34 pameyer
no hurry from my end
18:34 pdurbin
Did you exercise IngestableDataChecker?
18:34 pameyer
I assumed that if I broke that the integration tests would tell me
18:34 pdurbin
ok
18:35 pameyer
... but I don't know for a fact that there's coverage, and didn't check it
18:35 pdurbin
I like how you're using "try with resources" there anyway.
18:36 pdurbin
pameyer: oh, the other thing is that you're putting tabs everywhere. :)
18:37 pameyer
pdurbin: the tabs may be completely accidental
18:42 pdurbin
pameyer: ok, I think I'm done. I made two commits. I'm going to leave it to you to back out of changes to BatchServiceBean and FileRecordJobListener. Again, my rule of thumb is that if I didn't exercise the code and I'm not sure how to instruct QA to exercise the code, I don't make the change.
18:43 pameyer
pdurbin: ok - I'll back those 2 out, and re-check to make sure that they weren't actually doing something
18:52 pdurbin
pameyer: how are we looking with DCM S3? Should I go ask Kevin?
18:54 pameyer
pdurbin: I think it's pretty close, but Kevin probably has more info than me
18:54 pdurbin
pameyer: ok. He hasn't asked you to make any more changes to the branch or anything?
18:56 pameyer
not since the AWS CLI changes
18:57 pdurbin
ok, thanks
19:03 jri joined #dataverse
19:57 pdurbin
pameyer: sounds like you should check in with Kevin about the DCM S3 stuff
19:58 pameyer
pdurbin: more recently than ~10m ago?
19:59 pameyer
we were just talking on slack
20:15 pdurbin
if you're talking on slack, that's good
21:39 pameyer left #dataverse
23:04 jri joined #dataverse
