Time
S
Nick
Message
07:01 poikilotherm joined #dataverse
11:59 donsizemore joined #dataverse
12:00 donsizemore
@pdurbin more to fix in the ansible role but, as with my housekeeping, a lick and a promise
12:01 pdurbin
donsizemore: well, your most recent push unblocked me so thanks!
12:04 poikilotherm
Ah, 8am @ Boston... :-D Good morning guys. :-)
12:09 pdurbin
I could use a cup of coffee. And I'm planning on finishing some dishes and going to the gym on my way to work.
12:10 pdurbin
Maybe I'll just make some coffee here. Good morning.
12:16 pdurbin
donsizemore: today I might try the extra args thing. I hope it works.
12:18 pdurbin
poikilotherm: thanks for catching that jacoco/coveralls thing
12:18 poikilotherm
Sure, you're welcome :-)
12:19 poikilotherm
Browsed throught Travis output and saw that the reports failed (which is not failure to the build)
12:19 pdurbin
Ah. Makes sense.
12:19 pdurbin
Probably weeks or months would have gone by before I noticed.
12:21 pdurbin
Coveralls is only showing unit tests. I'd love to have coveralls or some other website show integration test coverage over time.
12:29 poikilotherm
While working with OpenClover, I combined the results of surefire and failsafe... Maybe this could be a solution for Coveralls, too?
12:30 poikilotherm
And of course you could think about switching to OpenClover, which has plugins for Jenkins, too.
12:30 poikilotherm
I used that on GitLab CI and it worked well, too.
12:31 poikilotherm
(But of course there was no nice and fancy integration to show the html report - had to release those as artifacts)
12:32 poikilotherm
And you would have to switch to use failsafe, too... (Which is a good idea anyway)
12:33 poikilotherm
(Switch = use failsafe for IT, surefire for UT)
12:35 pdurbin
I'm fine with switching to whatever but I really like how Coveralls shows the percent change in code coverage over time.
12:38 poikilotherm
You can definitly do that in OpenClover, but of course it is not as easy as "just use the coveralls maven plugin". Using such a service is very comfortable... ;-)
12:41 poikilotherm
Maybe codecov.io is worth a try? Never used...
12:43 poikilotherm
They offer merging of reports in their docs. No idea how good this works
12:56 pdurbin
Sure, maybe we could continue to use Coveralls for unit tests and start using codecov.io for integration tests. It's used by this IQSS project: https://codecov.io/gh/IQSS/Zelig
12:56 pdurbin
Anyway, I'm about to hop on my bike to go to work.
12:57 poikilotherm
Oh, I thought you were already there... Read you later... ;-)
14:24 pameyer joined #dataverse
14:27 poikilotherm
Hey guys, is @landreev still with you and available here on IRC ?
14:28 poikilotherm
Or is this someone outside Harvard? It is not clear from the Github profile...
14:30 pameyer
@poikilotherm - landreev was still at iqss yesterday :)
14:31 pameyer
he's on the iqss dev team - I haven't seen him in irc much
14:31 poikilotherm
Ok, then I'll open an issue and mention him there
14:31 poikilotherm
Thx
14:34 pameyer
np
14:35 pameyer
pdurbin: any tips on places in the ITs that store a setting value, change it for tests, and reset to the original value?
14:54 pdurbin
poikilotherm: he's in Slack
14:56 pdurbin
pameyer: maybe look at SearchApiNonPublicAllowed in @BeforeClass and @AfterClass of SearchIT.java
14:58 pameyer
pdurbin: thanks, I'll take another look (I ended up in SearchIT for changing setting values)
15:05 pdurbin
pameyer: sure, please just lemme know if you have questions
15:07 pameyer
pdurbin: are you sure that SearchIT stores the value of the setting that was there before it runs?
15:11 pdurbin
I'm not sure. I'm not sure what you mean.
15:11 pdurbin
Oh, you want to put back what was there. I wouldn't worry about that.
15:12 pdurbin
The harvesting server is off by default. You should just turn it back off when you're done with the test, I think.
15:12 pdurbin
Thanks for working on this.
15:13 pameyer
well, I run them in a fresh container; so it wouldn't be a problem for me
15:13 pameyer
but since they passed for landreev before his PR, he might be defaulting to on
17:24 donsizemore joined #dataverse
17:54 donsizemore joined #dataverse
18:41 pdurbin
pameyer: xmas came early for you: https://groups.google.com/d/msg/dataverse-community/TlQPNI3Ip2E/hGr63CaDFwAJ
18:44 pameyer
pdurbin: haven't kicked the tires yet; but yeah
18:45 pameyer
I'll take being annoyed over node.js vs being annoyed that I can't do bulk offline validation :)
18:47 pdurbin
:)
18:48 * pdurbin
runs ./ec2-create-instance.sh 5067-file-type-facet-caps
18:48 pdurbin
donsizemore: ^^
18:49 donsizemore
@pdurbin can i be one of the authors on the paper!
18:50 pdurbin
heh. of course!
18:50 pdurbin
Actually, I think this is a fine way to expose Dataverse developers to all the great work you've been doing with Ansible.
18:51 donsizemore
it may atone for my years of stupid questions
18:51 pdurbin
heh
18:51 pdurbin
donsizemore: please remind me if you expect non-IQSS repos to work or not
18:51 donsizemore
@pdurbin i don't see why they shouldn't
18:51 pdurbin
cool
18:52 pdurbin
And extra vars? Do we expect it to work? If so, how would I pass the GitHub org name (OdumInstitute or whatever) to Ansible?
18:54 pameyer
donsizemore - I think you've been more effective spreading the gospel of ansible than I have
18:54 pdurbin
I *hear* pameyer uses Ansible. You gotta have faith! :)
18:55 pameyer
paranoia, and our ops ansible repo being private :)
18:55 pameyer
besides - I learned from the DCM that "use ansible" wasn't a great installation experience for non-developers ;)
18:56 pdurbin
yeah, let's stick with Perl ;)
18:57 pameyer
it occured to me a while back that I should make a PR on dataverse-ansible to make it imdepotent (re-runnable when/if parts failed, since I probably spelled that wrong)
18:57 pameyer
but that's currently languishing on my todo list
19:00 pameyer
pdurbin: eventually the installer will get rewritten in something
19:02 pdurbin
Tcl
19:09 pdurbin
pameyer: do you know the syntax for extra args?
19:09 pameyer
for ansible? not offhand
19:10 pdurbin
Ok. I guess I'll google it. donsizemore: to pass the non-IQSS GitHub org
19:12 pdurbin
oh, right... --extra-vars "dataverse.git.branch=9999-get-my-margarita"
19:12 pdurbin
vars not args
19:18 * pdurbin
crosses fingers
19:48 Mahsa joined #dataverse
19:50 Mahsa
Hi, I just fixed my glassfish -> Apache fronting. Pameyer shared virtual server tages from domain.xml here
19:50 Mahsa
http://irclog.iq.harvard.edu/dataverse/2018-09-18#i_72659
19:50 Mahsa
What I noticed was that jk-connector was missing from my virtual serveres list, depsite the fact that it was enabled on http-listener1
19:51 Mahsa
Thought I should report here for future use. Thank you all.
19:51 pdurbin
interesting, thanks Mahsa
19:51 pdurbin
I'm glad you got it working!
19:52 Mahsa
Thanks!
19:52 Mahsa left #dataverse
19:55 pameyer
Mahsa - glad to hear that you got it working
20:01 pameyer joined #dataverse
20:19 pdurbin
pameyer: oh man, I'm breaking Don's toys: https://travis-ci.org/IQSS/dataverse-ansible/builds/431205138
20:24 pameyer
pdurbin: do you know about yq?
20:24 pdurbin
yep, just found and installed it
20:27 pameyer
one of ansible's fiddly parts is "there's a syntax error somewhere, that might or might not be around here"
20:28 pdurbin
Ok. I'll keep that in mind. Thanks.
20:35 pdurbin
nice, got --extra-vars working
20:35 pdurbin
made a pull request: https://github.com/IQSS/dataverse-ansible/pull/27
20:47 pameyer
puzzeling - no var_files in the role definations
20:47 pameyer
guess I'm out of date
20:48 pdurbin
¯\_(ツ)_/¯
20:53 donsizemore joined #dataverse
20:53 donsizemore
hey @pdurbin you still around?
21:28 pdurbin_m joined #dataverse
21:28 pdurbin_m
donsizemore: sort of. any thoughts on https://github.com/IQSS/dataverse-ansible/pull/27 ?
21:29 donsizemore
@pdurbin_m it fixes your use case, but breaks a vanilla run. can you pass it 'dataverse.git.branch' and 'dateverse.git.repo'?
21:30 donsizemore
the values in defaults/main.yaml should be superceded by the extravars flags
21:31 pdurbin_m
donsizemore: I tried. didn't work.
21:33 donsizemore
you linked to the 2.4 docs, but for all 2.n versions the -e value should win: "extra vars (always win precedence)"
21:35 pdurbin_m2 joined #dataverse
21:37 pdurbin_m2
donsizemore: right, I get that it should win, but it doesn't. In the docs I linked to do you see all the {{examples}} in curly braces?
21:37 donsizemore
correct
21:37 pdurbin_m2
so I added them. and now it works
21:38 pdurbin_m2
doesn't work without curly braces
21:38 donsizemore
as long as a user not launching the playbook from a script edited the group_vars or passed the extra vars it should be fine
21:39 donsizemore
but one of the early requirements per bencomp was a standalone role which would produce a current, default dataverse installation with minimal fuss on the part of the user
21:39 donsizemore
just curious, which version of ansible are you calling (definitely 2.4, or 2.6?)
21:41 pdurbin_m2
2.4.2.0 I think
21:41 donsizemore
any chance the script could update to 2.6 (provided by epel) before calling the playbook? would be great to stick with the --extra-vars
21:41 pameyer
that feels like it's doing templating in the varible definations, instead of in the tasks/jobs/definations/etc
21:41 pdurbin_m2
typing with thumbs, picking up my kid from school
21:42 donsizemore
this can definitely wait! but maybe we can try updating ansible? it's under heavy development
21:42 pdurbin_m2
donsizemore: wanna try the script? :)
21:43 donsizemore
sure thing, i'm looking at the PR to get back to your scripting work
21:45 pdurbin_m2
nice. I'm pretty happy with how the script is shaping up
21:45 donsizemore
this is ec2.sh? we've got AWS access through another group here on campus
21:47 pdurbin_m2
yeah
21:48 pdurbin_m2
donsizemore: https://github.com/IQSS/dataverse/commit/be995cc
21:48 donsizemore
i'll definitely take a look. i may have questions (such as, what launches the installer?)
21:48 donsizemore
aha
21:52 donsizemore
centos7 should pick up ansible-2.6 from epel. i bet it's already installed and a previous version.
21:53 donsizemore
i'd try an additional "yum -y update git ansible" after line 76 as an easy test
21:53 donsizemore
i'd like to head for home as well, and maybe pick this up in the morning? i'd love to get it working
21:54 pdurbin_m2
donsizemore: docs: https://github.com/IQSS/dataverse/commit/0ae6dc186d9f00eaa76ce305cc196ace980a3cfc
21:54 pdurbin_m2
donsizemore: manana! go home!
21:55 donsizemore
https://www.youtube.com/watch?v=8N_tupPBtWQ
21:58 pdurbin_m2
classic
21:59 pdurbin_m2
pameyer: you might like this script too
22:21 pameyer left #dataverse
