Time
S
Nick
Message
04:09 LavetteGeneratio joined #dataverse
04:26 LavetteGeneratio left #dataverse
07:33 poikilotherm joined #dataverse
07:48 jri joined #dataverse
08:19 jri_ joined #dataverse
08:36 stefankasberger joined #dataverse
08:38 stefankasberger
@pdurbin @poikilotherm Regarding your request about k8s. Never done anything with k8s so far, so its definitely not my/our cluster. :)
08:49 poikilotherm
Hi stefankasberger :-)
08:52 poikilotherm
Hmm I had another look at the VMs and I don't know where pdurbin saw your name on it...
08:53 poikilotherm
However - I just stopped the VMs. So they are not deleted, in case someone needs them ;-)
08:57 jri joined #dataverse
09:07 YSF23 joined #dataverse
09:07 YSF23
Hello everyone. do you have a trick to stay connected to the channel please
09:07 poikilotherm
HI YSF23
09:08 poikilotherm
Connected to what channel? This IRC channel?
09:13 YSF23
only here for chat
09:14 poikilotherm
Are you using a client or using the web client at dataverse.org?
09:15 YSF23
no no
09:31 YSF23
But it does not matter, I have a question for the update.
09:31 YSF23
I have version 4.11 and I would like the last
09:32 YSF23
You have a trick , please ?
09:35 poikilotherm
Well you will have to work to 4.12 first, which introduced Flyway support.
09:36 poikilotherm
From there, you can move to the latest release, but scan the release notes for things to do, as some things changed and might need a reindex etc
09:36 poikilotherm
Better safe than sorry: do backups of your database and files before doing the upgrades.
09:37 YSF23
I have to do 4.11 and then 4.12 -> 4.13 -> 4.14 .....?
09:37 poikilotherm
No.
09:37 poikilotherm
You will have to go from 4.11 to https://github.com/IQSS/dataverse/releases/tag/v4.12
09:38 poikilotherm
Then you should be able to upgrade to https://github.com/IQSS/dataverse/releases/tag/v4.17 and skip those in between
09:38 YSF23
Ah oke, thank you very much :)
09:38 poikilotherm
But as some things changed in those releases in between 4.12 and 4.17, you should read those release notes
09:39 poikilotherm
And look for things that might be important for you
09:39 poikilotherm
And most likely you will have to do a reindex and reexport
09:39 poikilotherm
Please create backups before
09:39 YSF23
Okay, I'll read this first
09:43 juancorr joined #dataverse
10:43 stefankasberger
Does anyone know in here, how it worked in older dataverse versions, to change the piwik tracking code?
10:43 stefankasberger
There is documentation about it, but the curl commands alone do not work so far. http://guides.dataverse.org/en/4.6.2/installation/config.html#googleanalyticscode
10:43 stefankasberger
There is documentation about it, but the curl commands alone do not work so far. http://guides.dataverse.org/en/4.6.2/installation/config.html#piwikanalyticsid
10:59 pdurbin
stefankasberger: maybe it would help to look at the pull request where piwik was added: https://github.com/IQSS/dataverse/pull/3374
11:03 pdurbin
YSF23: the trick I was thinking about is for installations that are way back on 4.8 or whatever: http://irclog.iq.harvard.edu/dataverse/2019-10-23#i_110230 ... But you're on 4.11 so this trick doesn't apply to you. Like poikilotherm says, once you're on flyway (4.12) you should be able to do some skipping.
11:05 YSF23
Yes I thought thank you very much for your help
11:05 poikilotherm
Moin moin ihr Landratten :-) (Good morning you landsmen)
11:08 pdurbin
YSF23: I'm going to Belgium for the first time in early February. I'm pretty excited about it. :)
11:10 YSF23
Great, I welcome you. If you come to brussels I can welcome you
11:10 YSF23
You have a instagram or a other ?
11:23 stefankasberger
@pdurbin: unfortunately, the PR does not help.
11:24 stefankasberger
I guess there is always a glassfish restart necessary after changing the settings or the piwik xhtml file. Or is there something else to do?
11:34 pdurbin
stefankasberger: all that code was refactored (as you know) so maybe you should upgrade Dataverse first and turn on piwik afterwards? I've never actually set up piwik (or google analytics) so I'm not sure what the gotchas are, if any.
11:42 YSF81 joined #dataverse
11:49 pdurbin
YSF81: yes, I'll be in Brussels! For FOSDEM. And I'm not on Instagram but I am on Twitter: https://twitter.com/philipdurbin
11:49 pdurbin
The dates and details are in the "Dataverse Meetups" spreadsheet linked from the bottom of https://dataverse.org/events
11:52 donsizemore joined #dataverse
11:57 donsizemore
@stefankasberger I set up Google analytics (not Piwik) but the HTML snippet sat outside the Glassfish root so no restart was necessary
11:57 donsizemore
(I think?)
12:06 stefankasberger
thanks. will try that.
12:10 YSF81
Here I add you on twitter, if you need help when you are here do not hesitate :)
12:13 pdurbin
YSF81: thanks! I just followed you back. Any plans to attend FOSDEM? I submitted a talk about Dataverse. :)
12:16 YSF74 joined #dataverse
12:17 pdurbin
donsizemore: morning. I left you a comment at https://github.com/IQSS/dataverse/pull/6316#issuecomment-547128433 . I'm happy to cherry pick stuff and otherwise mess with your branch if you want.
12:17 donsizemore
@pdurbin i want to talk to akio to make sure i won't screw up his TRSA fork
12:18 pdurbin
ok, no rush
12:18 donsizemore
@pdurbin in hindsight i should've test from sub-branch against sub-branch instead of our develop. oh, well.
12:18 donsizemore
@pdurbin but at the end of the day I just want to add those two files
12:18 pdurbin
Right. I'd say you could just make a fresh branch off develop and add them.
12:19 YSF74
It's really interesting, I will of course come
12:19 pdurbin
YSF74: great! My talk hasn't been accepted yet.
12:19 pdurbin
poikilotherm: I'm also thinking about submitting a talk about my side project. :)
12:20 poikilotherm
Nice!
12:20 donsizemore
@pdurbin I did that, but it included all the commits to develop
12:20 donsizemore
@pdurbin how does Kevin feel about GitHub's built-in Squash-and-Merge button?
12:20 pdurbin
donsizemore: hmm, maybe re-clone? That's what I do. :)
12:20 donsizemore
@pdurbin the commits are in our fork
12:21 YSF74
Okay, I'd be happy to meet you :)
12:21 pdurbin
Right, I think you should make develop on Odum the same as develop on IQSS. I should probably work from a fork more often to feel this pain. Also, I don't think kcondon has ever used that new button.
12:23 poikilotherm
donsizemore pdurbin: that's why I use the pull automation from github
12:23 pdurbin
poikilotherm: but don't you get extra merge commits?
12:23 poikilotherm
https://github.com/IQSS/dataverse/issues/6053
12:23 poikilotherm
Nope
12:24 poikilotherm
At least not in develop
12:24 poikilotherm
My develop follow upstream 1:1
12:24 pdurbin
oh, good
12:24 poikilotherm
Don't touch it... Otherwise you will have 'em in there ;-)
12:24 poikilotherm
I also wanted that for master branch, but my proposal has been rejected...
12:25 pdurbin
maybe you can sell donsizemore on it
12:25 pdurbin
I prefer a more manual approach.
12:25 poikilotherm
Sure
12:25 donsizemore
our fork's develop is current with IQSS, that's not the ish
12:26 pdurbin
donsizemore: oh! So maybe just re-clone? I'm confused.
12:26 donsizemore
you don't want all my commits from testing the automated PR build, and without throwing Akio's TRSA branch for a loop I think the easiest thing to do is select Squash-and-Merge instead of Merge
12:26 donsizemore
to undo all those testing commits manually i have to rebase, and i don't want to mess Akio up
12:28 poikilotherm
Well I had this situation a few times.
12:28 poikilotherm
There are multiple ways out
12:29 poikilotherm
Shouldn't using a new branch solve this for you?
12:29 pdurbin
donsizemore: I guess I'm confused why we can't just copy those two files to a new branch off develop. I'm happy to do that if you want.
12:30 donsizemore
@poikilotherm I still picked up all the commits in develop on our fork
12:30 pdurbin
Maybe you're trying to preserve some of the history of changes to those two files along the way?
12:30 donsizemore
@pdurbin that will be fine for this PR but I'll have the same situation next time
12:31 poikilotherm
Ah so you don't have a clean develop branch from upstream
12:31 poikilotherm
But a mixture of IQSS develop and your own commits to it?
12:31 donsizemore
correct - i was testing the PR stuff against our fork's develop, thinking I'd just change "OdumInstute" to IQSS in the Jenkins job when the time came
12:32 pdurbin
I think the problem is that there are two divergent "develop" branches in play. I would recommend that forks don't use "develop" or "master" if possible and if that's reasonable.
12:32 donsizemore
and Akio is maintaining a TRSA branch off our fork that I don't want to git-bomb
12:32 poikilotherm
Ok I see two options. Create a new fork and use that for a new PR.
12:32 donsizemore
@pdurbin in hindsight, absolutely
12:32 poikilotherm
Or go ahead, create a branch from your develop and reset to a common ancestor
12:32 poikilotherm
then merge in upstream develop
12:32 donsizemore
GitHub also boasts a Merge-and-Squash button
12:33 poikilotherm
and create a new commit for your changes
12:33 poikilotherm
Or do a rebase of the new branch to upstream - you will not destroy your develop branch that way
12:37 donsizemore
hmm, I may try that. I'm not losing sight of the fact that I've done weeks of worth for free, and can just as easily close the PR, though.
12:37 donsizemore
^^work
12:38 poikilotherm
Well that's sth. pdurbin, djbrooke and others at IQSS need to address.
12:39 poikilotherm
I'd be happy to help whatever you want me to help with :-)
12:39 poikilotherm
+you
12:59 pdurbin_m joined #dataverse
13:01 pdurbin_m
donsizemore: yes, weeks and weeks of free labor. Thank you! I'm sorry that my interest in got hygiene is causing trouble. I can ask kcondon to try that button if you'd like.
13:01 pdurbin_m
git* hygiene
13:02 donsizemore
@pdurbin_m i normally don't work off develop, but stupidly did so on our fork testing the PR stuff... if it weren't for akio's TRSA fork i'd just rebase
13:08 pdurbin_m
Im at the gym but I'll look closer in a bit. Thanks. And sorry again for all the trouble.
13:13 poikilotherm
donsizemore: just give me a shout out if you need any help. There is also "rebase -i", which allows you to not squash everything, in case you don't want to purge all the history and your work.
13:13 donsizemore
@poikilotherm it's not the history i'm worried about, it's akio's branch off our fork. of course, he has to merge it against develop each time there's a new release anyway, so maybe it's a non-issue
13:14 poikilotherm
Well you can't really damage much stuff... As long as you don't push to Github, you can always go back on your local machine to where you came from
13:15 poikilotherm
So if you screwed up local, just reset to origin ;-)
13:23 donsizemore
i was testing the PR stuff in jenkins though, i had to push
13:23 donsizemore
and i wouldn't be me if i did anything right ;)
13:25 poikilotherm
No no, it has been the same for me.
13:26 poikilotherm
What I did to avoid these "trigger pushes": I created a branch from my feature branch, added them there and fired of to Jenkins
13:26 poikilotherm
That way I hadn't to clean 'em out everytime
13:43 pdurbin
Ok, I'm at my desk. Should I start looking at branches?
13:43 donsizemore
@poikilotherm what i should've done in hindsight...
13:43 donsizemore12 joined #dataverse
13:44 donsizemore12
software demo in 15. back in a bit!
13:44 donsizemore12 left #dataverse
15:36 stefankasberger
So, the change to our new Matomo instance works now. :)
15:36 stefankasberger
Another Matomo related question: who here uses Matomo too, to analyze visitors to Dataverse?
15:37 stefankasberger
We use it for Dataverse Production, and also our Typo3 Website. So I am asking myself, if cross-domain-linking would make sense (to track visitors across domains) or to use the actual way, to setup 2 websites - one for each (dataverse, typo3) and track them independently.
15:41 pdurbin
Well, https://github.com/IQSS/dataverse/pull/3374 was created by jri so I assume he uses piwik/matomo. :)
15:43 jri
Yes, we're using it too :)
15:44 jri
but we don't have a cross-domain-linking use case unfortunately :(
15:44 jri
I can't talk about this behaviour
15:45 * poikilotherm
goes hunting for kids.
15:45 poikilotherm
People might enjoy taking a sneak peek at https://dataverse-k8s.readthedocs.io/en/latest/day1/config.html
15:45 poikilotherm
Cu around
16:27 jri joined #dataverse
19:41 poikilotherm joined #dataverse
19:42 poikilotherm
Good evening ladies and gentlemen
19:48 pdurbin
poikilotherm: nice docs
19:50 poikilotherm
Thx
19:50 poikilotherm
I'm really trying to get issue 4665 (docker in prod) done here ;-)
19:51 poikilotherm
Maybe once this is done, we should refactor the openshift/... part of the "installation guide" to just point to this new guide ;-)
19:52 pdurbin
poikilotherm: sure, but not just the openshift part. Blank out the whole page and just link to your guide. :)
19:52 poikilotherm
Aye
19:53 poikilotherm
And it should not be hidden in the dev guide only :-D
19:54 pdurbin
Oh. Good point. Where should it go?
19:54 poikilotherm
In both Installation and Dev
19:55 poikilotherm
"Developemnt Environment" seems to be a good point to link to these docs for using it in dev mode
19:56 pdurbin
I'd agree but we are trying to keep that page short.
19:56 pdurbin
For the Installation Guide, there's a "choose your own adventure" section you can add to.
19:56 poikilotherm
Sure. Just a link like: "instead of doing all of the below, you might instead opt for using ..."
19:57 poikilotherm
Yeah, that would be definitly a good place to start in the installation docs
19:57 poikilotherm
Some other places might be nice, too - it's just a hint that there are other options, too
19:57 poikilotherm
No need to do exhaustive descriptions, that's what the K8s guide is for
20:01 poikilotherm
I was thinking about giving the docs a name containing "guide", so it is more inline with the others
20:01 poikilotherm
Any suggestions for a name?
20:01 poikilotherm
Just Kubernetes seems to be a bit shortened, as I do have ideas about using this stuff on Docker only, too
20:07 pdurbin
poikilotherm: sorry, was troubleshooting a new Netbeans installations. What's that?
20:08 pdurbin
You could make a whole new guide if you want.
20:08 pdurbin
a new top level guide, I mean
20:08 poikilotherm
I am already :-D
20:09 pdurbin
well, I mean along side the other guides
20:09 poikilotherm
I dunno if we can create a link to the k8s docs at toplevel
20:09 poikilotherm
Yeah
20:09 poikilotherm
That'd be cool
20:09 pdurbin
Oliver's Guide
20:09 poikilotherm
But how should we call it
20:09 poikilotherm
...
20:09 poikilotherm
That's not helpfull, but funny
20:09 pdurbin
Oliver's Rough Guide to Dataverse
20:10 poikilotherm
The Kubernetes Guide to Dataverse
20:10 poikilotherm
It's 42
20:10 pdurbin
Dataverse Survival Guide
20:10 poikilotherm
What about "Dataverse Cloud Guide"?
20:11 poikilotherm
Maybe even crazy "DataverseCloud Guide"
20:11 pdurbin
ending with "Guide" is good
20:11 poikilotherm
Like DataverseTV etc
20:12 poikilotherm
There is Solr and SolrCloud
20:12 pdurbin
Container Guide
20:13 poikilotherm
Cloud & Container Guide?
20:17 pdurbin
I think I'm a little too spent to decide right now.
20:18 pdurbin
maybe just link to your external guide for now
21:05 poikilotherm
Yeah... I just made docker compose work with dataverse-kubernetes
21:36 poikilotherm
https://github.com/IQSS/dataverse-kubernetes/issues/118
21:36 poikilotherm
Easier than I thought...
21:36 poikilotherm
We'll see if I can make Slava happy :-D
21:43 pdurbin
This is for working on Dataverse code itself too? Or is it for configs and stuff?
21:48 poikilotherm
I build the dev image, so you can use any arbitrary branch of Dataverse in it
21:50 pdurbin
How do you switch branches?
21:51 poikilotherm
Put code in dataverse directory (or better: use the submodule). Use git as used too
21:51 poikilotherm
(Within that directory)
21:52 pdurbin
Ok, so I'm not limited to "dataverse @ ee874a6" which is where the submodule points.
21:54 poikilotherm
Nope.
21:54 poikilotherm
It's just a submodule.
21:54 poikilotherm
You can do whatever you want in it
21:54 pdurbin
After a build an image for an arbitrary branch, do I have to push that image to Docker Hub to use it?
21:54 pdurbin
I* build
21:55 poikilotherm
Nope
21:55 poikilotherm
Neither with k8s nor docker-compose
21:55 pdurbin
Ok, it just shows up under `docker images` or something.
21:55 poikilotherm
Aye
21:55 pdurbin
What's the name of the image?
21:56 poikilotherm
It gets tagged as `iqss/dataverse-k8s:dev` locally
21:56 poikilotherm
At least when using docker-compose
21:56 poikilotherm
Using skaffold, it gets a random ID
21:56 poikilotherm
Sry, no docs yet :-D
21:57 poikilotherm
This is just a PoC... ;-)
21:57 pdurbin
Ok. What's the advantage over what I do now? Right now I run Glassfish and Postgres and Solr on my Mac.
21:59 poikilotherm
It does everything for you
21:59 poikilotherm
No need to manually install anything
21:59 poikilotherm
And you can very easily test things
22:00 poikilotherm
Like "let me check if this runs on Postgres 10"
22:00 poikilotherm
Just be swapping out the image tags
22:00 poikilotherm
s/be/by/
22:00 poikilotherm
No need to fiddle with Glassfish installation.
22:01 poikilotherm
And by using JRebel, reload times are ~ zero.
22:01 pdurbin
How long from when I change a line of Java code until when I can see the change with curl or a browser?
22:01 poikilotherm
With JRebel? A few seconds.
22:01 pdurbin
I don't have JRebel.
22:02 poikilotherm
Yeah. Then you'll have to go through a full deployment. Or you can use the Java debugger, which can also change code live.
22:02 poikilotherm
But only very limited
22:02 poikilotherm
No new signatures etc
22:02 poikilotherm
And this does not help you with JSF :-/
22:03 poikilotherm
But recently I think I understood that there are some changes that break that for you...
22:03 pdurbin
Sorry, what changes?
22:03 poikilotherm
Wasn't there some Netbeans/Jacoco/... things that lead to a non-working hot reload for JSF resources?
22:04 pdurbin
Oh, that. Well, we have a workaround. Kind of a stupid workaround but we could improve on it.
22:04 poikilotherm
Yeah, we talked about that
22:04 pdurbin
You have JRebel? How much does it cost?
22:05 poikilotherm
Yeah, I bought a license. Wanted to report a success story later on. Had no time for that yet. Let me check the invoice.
22:06 pdurbin
That's ok. I was just wondering if you knew the cost off the top of your head.
22:07 poikilotherm
450 bucks
22:07 pdurbin
yikes
22:07 poikilotherm
?
22:07 pdurbin
That's a lot for a vi user. :)
22:07 poikilotherm
LOL
22:07 poikilotherm
But it works like a charm
22:07 poikilotherm
For everyting
22:08 poikilotherm
JSF, code, ...
22:08 poikilotherm
Nice IDE integration
22:09 pdurbin
like dev mode in Quarkus :)
22:09 poikilotherm
Aye
22:09 pdurbin
I gotta get my kid from her cello lesson down the hall, but I'll leave you all with this: https://github.com/oblassers/dmap/issues/1#issuecomment-547553329
22:10 poikilotherm
I'm really glad we some project money left for this. That was a great speedup for developing.
22:10 poikilotherm
:-D
22:10 poikilotherm
Cu tomorrow
