Time
S
Nick
Message
11:52 poikilotherm joined #dataverse
14:14 djbrooke joined #dataverse
14:44 djbrooke joined #dataverse
14:46 donsizemore joined #dataverse
14:50 Jim_ joined #dataverse
14:51 djbrooke joined #dataverse
14:54 Jim_
hi all - I was just going through a test 4.8.6 to 4.9.almost3 upgrade on TDL's dev machine (looks good so far) and wondered if anyone has thought about amending the 4.9 install instructions to point out you should not do registerDataFileAll if you're going to turn file PIDs off in 4.9.3?
14:55 pdurbin
Jim_: oh, the release notes for 4.9?
14:55 Jim_
yes
14:56 Jim_
perhaps obvious, but a bad thing to do if you wrotely update version by version
14:56 Jim_
rotely
14:56 pdurbin
yeah
14:57 pdurbin
release notes aren't versioned the way guides are
14:57 poikilotherm joined #dataverse
15:00 pdurbin
Jim_: I just added a warning. Please let me know what you think: https://github.com/IQSS/dataverse/releases/tag/v4.9
15:11 djbrooke joined #dataverse
15:11 djbrooke joined #dataverse
15:16 poikilotherm
Hi pdurbin, could you give me hint where I find the code that runs tasks at startup of dataverse and where I could find something about the default workflows being initially deployed to a new install?
15:17 poikilotherm
I'm a bit lost in the code base :-D
15:23 Jim_
pdurbin - looks good except for the type (what instead of want). I guess it could be even more explicit - if you are upgrading to 4.9.3 or beyond and intend to disable file DOI creation, do not do this step.
15:27 Jim_
As for workflows, the ones I know of are all just examples - in scripts/api/data/workflows and there are no workflows enabled by default. http://guides.dataverse.org/en/latest/developers/big-data-support.html has some info and the api to check on them is described on http://guides.dataverse.org/en/latest/api/native-api.html .
15:27 Jim_
(just been working to add a post-publication workflow to submit archival copies of data to the Digital Preservation Network...)
15:36 djbrooke joined #dataverse
15:38 djbrooke joined #dataverse
15:42 djbrooke joined #dataverse
15:58 poikilotherm joined #dataverse
15:59 poikilotherm
@Jim_ that is the intention I am thinking in - but for me this would be related to PID minting at different stages in a very configurable way.
16:00 poikilotherm
I hoped the Workflows would be the thing to look at for the different stages of a dataset from creation over update to publication
16:14 pdurbin
Jim_: thanks, I just changed it to this: **Please note: Do not run the `registerDataFileAll` command below if you do not plan to give your files persistent identifiers, which are no longer required in 4.9.3 or later (#4929).**
16:15 pdurbin
poikilotherm: hi, so you have two questions? Tasks at startup? Docs on workflows?
16:15 poikilotherm
I found the workflows and how they are triggered thank to Jim_
16:16 poikilotherm
Currently digging into this...
16:16 poikilotherm
Maybe this is a good place for the PID minting stuff
16:16 poikilotherm
The other question is about startup tasks
16:17 poikilotherm
Do you have such a construction in place?
16:17 poikilotherm
E. g. to validate configs?
16:17 poikilotherm
I'm asking because that could be usefull to load some kind of plugin for the PID providers
16:19 pdurbin
poikilotherm: I would suggest looking at the @PostConstruct startup method that calls registerProviderFactory at https://github.com/IQSS/dataverse/blob/v4.9.2/src/main/java/edu/harvard/iq/dataverse/authorization/AuthenticationServiceBean.java#L111
16:20 pdurbin
there's also something for starting the EJB timer stuff
16:22 poikilotherm
Ah thx
16:22 poikilotherm
Will look into that.
16:22 pdurbin
sure
16:22 poikilotherm
I forgot: an update about #4106, the pid plugin issue...
16:23 poikilotherm
I have notice from Felix Bensmann (fbgesis) that he is not involved in this stuff anymore
16:23 poikilotherm
There is another employee from GESIS following up to his work, but I hadn't a chance to talk to him yet.
16:23 poikilotherm
Will keep you posted how it goes.
16:26 pdurbin
huh, are you sure he isn't involved anymore?
16:29 poikilotherm
Yes - he wrote me an email today :-D
16:29 poikilotherm
I can forward this to you, but it is in german... ;-)
16:30 poikilotherm
Ok, he said he is only doing "consulting" from now on - the actual implementation work seems to be done by a Mr. Klas
16:32 poikilotherm
https://www.gesis.org/institut/mitarbeiterverzeichnis/person/?tx_gextstaffdir_staffdirectory%5Bemail%5D=claus-peter.klas%40gesis.org&tx_gextstaffdir_staffdirectory%5Baction%5D=details&tx_gextstaffdir_staffdirectory%5Bcontroller%5D=Index&cHash=2a391c38fc0fa940c580b70a939365f7
16:39 pdurbin
ok, well, you're welcome to invite him to join us here
16:40 poikilotherm
Yes Sir! ;-)
16:40 poikilotherm
Hope to get in touch with him next week. We'll see.
16:42 pdurbin
cool
16:42 pdurbin
I'm glad you're thinking about the pluggable PID stuff.
17:16 poikilotherm joined #dataverse
17:16 poikilotherm
Sure. This is pretty urgent for us, as we need to get our repo in place more sooner than later...
17:19 poikilotherm
Do you think there is help wanted with the Java 11 stuff? This seems as it might blow up in the nearer future...
17:19 djbrooke joined #dataverse
17:22 pdurbin
You're welcome to see if you can reproduce the errors I'm seeing: https://github.com/IQSS/dataverse/issues/4259#issuecomment-425453563
17:44 poikilotherm joined #dataverse
17:53 poikilotherm joined #dataverse
19:07 poikilotherm
Hey @pdurbin, just updated #4259 with my experiences :-)
19:08 * pdurbin
looks
19:30 poikilotherm
While looking at the failing tests: have you guys ever thought about replacing jHove with something like Apache Tika?
19:30 poikilotherm
As far as I can see, jHove is only used for file type detection, right?
19:31 poikilotherm
Tika has quite a few more formats... ;-)
19:32 pdurbin
poikilotherm: Jim_ and djbrooke mentioned Tika at https://github.com/IQSS/dataverse/issues/5030
19:32 pdurbin
yes, I believe jhove is only used for file type detection
19:34 pdurbin
poikilotherm: did you see my comments on https://github.com/IQSS/dataverse/pull/5059 ?
19:36 poikilotherm
pdurbin: whooops - not yet. Thanks for the tip... *reading*
19:36 pdurbin
sure
19:42 pdurbin
poikilotherm: by the way, Jim_ is the one who wrote the google doc I was telling you about... his roadmap for stuff he's working on as a contributor
19:45 djbrooke joined #dataverse
19:45 poikilotherm
pdurbin: I thought so... ;-)
19:45 poikilotherm
pdurbin: Just commented your review :-)
19:46 pdurbin
thanks, I'm commenting on your comments
19:47 pdurbin
if it's easier, we can just talk here. I still have another 45 minutes before I need to pick up the kids
19:47 poikilotherm
As you prefer :-)
19:47 pdurbin
I like your proposed new names for the JVM options.
19:47 pdurbin
I think they help reduce potential confusion.
19:47 poikilotherm
Ah, the conflict results from a commit coming in after I rebased earlier this morning...
19:48 poikilotherm
Whatever you guys prefer... Would this trigger reiteration of reviews?
19:48 pdurbin
no, I don't think so
19:49 pdurbin
I mean, QA might find a bug and send it back to you.
19:49 poikilotherm
Ok, then I will change 'em...
19:49 djbrooke joined #dataverse
19:49 poikilotherm
Will take a few moments :-D
19:49 pdurbin
No rush. Thanks. Are you going to provide QA a Minio server to test with? That would be ideal.
19:50 poikilotherm
They can simply use the play.minio.io :9000 ;-)
19:50 poikilotherm
Credentials and stuff already in the docs :-D
19:51 djbrooke joined #dataverse
19:51 pdurbin
Ok. Maybe you can put that in the guide. I do something similar for Test Shib in the shib section of the guides.
19:51 pdurbin
"try out this test service to see if it works"
19:51 poikilotherm
About your comment regarding the empty compilerArgument - just leave a comment above it in pom.xml?
19:51 pdurbin
yes a <!-- comment -> please
19:51 poikilotherm
Sure.
19:52 poikilotherm
I will add the comment about the testing to the working combination report, ok?
19:52 pdurbin
working combination report?
19:59 poikilotherm
Yeah, new section
20:00 poikilotherm
"
20:00 poikilotherm
Reported Working S3-Compatible Storage
20:00 poikilotherm
Minio v2018-09-12
20:00 poikilotherm
Set dataverse.files.s3-path-style-access=true, as Minio works path-based. Works pretty smooth, easy to setup.
20:00 poikilotherm
HINT: If you are successfully using an S3 storage implementation not yet listed above, please feel free to open an issue at Github and describe your setup. We will be glad to add it here.
20:02 pdurbin
sure, sounds fine. and a link to the play thing
20:02 poikilotherm
Of course :-)
20:02 pdurbin
oh, you added it already
20:02 pdurbin
great
20:25 poikilotherm
DONE :-)
20:27 poikilotherm
And I'm done for today, too ;-) Its 22:27 here, let's call it a day.
20:28 poikilotherm
Have a nice weekend :-)
20:28 pdurbin
you too!
20:30 pdurbin
all of you
20:30 pdurbin left #dataverse
20:53 djbrooke joined #dataverse
20:58 djbrooke joined #dataverse
21:01 djbrooke_ joined #dataverse
21:30 djbrooke joined #dataverse
