Time
S
Nick
Message
07:03 jri joined #dataverse
08:12 jri joined #dataverse
08:20 stefankasberger joined #dataverse
08:30 stefankasberger joined #dataverse
08:31 stefankasberger
The Dataverse Conference proposal form does not work for me right now. Anyone else with this problem?
08:31 stefankasberger
https://projects.iq.harvard.edu/dcm2019/proposals
08:33 stefankasberger
The problem is two-folded: 1) I do not see the form on my Firefox 66.0.3 on Ubuntu (18.04 and 16.04). And on Chrome when I want to submit the filled out form, it tells me, the form can not connect with the server and i should check my connection.
09:53 pdurbin
stefankasberger: not sure, I'm on Ubuntu and the form looks ok on Firefox. I submitted a test from Chrome and it went through with "Thanks for your proposal submission! Well review the proposals and well contact you soon."
09:55 stefankasberger
Thanks. Have sent the proposal in an Email instead.
09:58 pdurbin
I saw you sent it to me and Danny. Thanks.
10:06 pdurbin
stefankasberger: your talk sounds great
10:06 stefankasberger
Thanks. :)
10:07 pdurbin
I was thinking about submitting one about improving communication throughout the project.
10:07 pdurbin
Last year I talked about integrations and eventually wrote them up at http://guides.dataverse.org/en/4.13/admin/integrations.html
10:08 pdurbin
Well, I wrote the initial page anyway. Others have contributed since then.
10:09 pdurbin
The problem is that https://dataverse.org/integrations isn't really kept up to date. Actually, now that I have edit access I should add a link to the guides from that page.
10:11 pdurbin
Done. Due to caching it might not show up for a while.
10:17 stefankasberger
Great. Have read the session description for preparation for mine. :)
10:18 pdurbin
Heh. "What did I say I'd talk about? Let me go read the description again." :)
12:28 donsizemore joined #dataverse
13:31 donsizemore
@pdurbin FWIW Dataverse deploys on PG 10 with sampledata, no SQL errors.
13:38 donsizemore
@pdurbin samesies with 11
13:59 pdurbin
nice!
13:59 pdurbin
I just left you a comment in the issue.
14:02 donsizemore
@pdurbin well, i'm recommending removal of 9.3, continued installation of 9.6 (unless you want to go for 10?) and updating to use postgres' new single-repo rpm
14:04 pdurbin
Huh. Can you please write more about the single repo rpm in the repo? And your thoughts on 10 or 11? I think we're happily on 9.6 for a while, using postgres as a service from AWS. For Harvard Dataverse. I'll drag it to code review so it comes up at standup.
14:05 donsizemore
@pdurbin write more, as a comment, or in the doc? https://yum.postgresql.org/repopackages.php sez "As of 15 April 2019, there is only one repository RPM per distro, and it includes repository information for all available PostgreSQL releases"
14:06 pdurbin
as a comment to be discussed at standup in about an hour, please
14:27 Phil joined #dataverse
14:28 Phil
test
14:29 pdurbin
Guest33381: hi
15:03 donsizemore joined #dataverse
16:50 donsizemore joined #dataverse
17:28 donsizemore
@pdurbin https://github.com/IQSS/dataverse-ansible/commit/5885b84b1b7624071fcf145fafa878c649ffc7b7
17:32 pdurbin
nice!
17:32 pdurbin
so allows for 11 too for when flyway supports it
17:33 donsizemore
11 deployed. no rpms just yet for 12
17:33 pdurbin
cool
17:33 donsizemore
so, i still need to tidy up letsencrypt (ugh) and then the API test suite
17:34 donsizemore
i think i can also publicly announce that jon successfully defended his dissertation proposal
17:34 pdurbin
wow! congrats to Jon!
17:35 pdurbin
if lets encrypt is no fun you could just merge what you've got. my bring your own cert use case worked fine
17:36 donsizemore
it's the letsencrypt-in-an-inaccessible-container case that's the bugaboo
17:36 pdurbin
because it's in vagrant?
17:36 donsizemore
vagrant doesn't let you expose ports below 1024 unless you run it as root
17:37 donsizemore
and lets encrypt wants 80
17:38 donsizemore
but the other cases (local certs, url-provided, pem-in-the-group_vars) all work locally for me
17:53 jri joined #dataverse
17:56 pdurbin
and maybe it works on ec2?
18:13 donsizemore
@pdurbin um, the plumbing is there, the dynamic DNS part isn't
18:18 pdurbin
I think I'm confused. I understand my "bring your own cert" use case. Is the other usecase to have a valid cert on a URL like https://ec2-3-81-171-108.compute-1.amazonaws.com ?
18:18 donsizemore
@pdurbin you'd need a cert for *.amazonaws.com (or specify your own hostname and pre-arrange some sort of DNS )
18:19 donsizemore
so you could bring your own *.dataverse.org cert and point a record at the instance once you spin it up
18:19 pdurbin
right. ok. so what's stopping us from merging it? we don't know how to do the dynamic dns stuff? can we do that part later?
18:19 donsizemore
i personally consider that out of scope.
18:20 pdurbin
me too
18:20 donsizemore
and if we drop the letsencrypt requirement from the issue, i think we should be good. or at least very close
18:20 pdurbin
so what's left that's in scope and can I help?
18:20 pdurbin
sure, let's drop the letscrypt requirment
18:21 pdurbin
the title is still "As someone spinning up Dataverse on EC2, I want valid SSL certs" ... and you've delivered on that! thanks!!
18:21 pdurbin
I could attach my forked version of the ec2-create script, if it helps
18:22 pdurbin
to https://github.com/IQSS/dataverse-ansible/issues/53 I mean
18:29 donsizemore
@pdurbin i'm covering my eyes and submitting it as a pull request
18:33 donsizemore
@pdurbin yay, merge conflicts! i'll sort them out. taking my car for an oil change this afternoon so it'll give me something to do at the dealership
18:34 pdurbin
heh
18:34 pdurbin
we can pick this up whenever
18:34 pdurbin
it's magic, I love it
18:34 pdurbin
did you get the "pem in mail.yml" to work?
18:35 donsizemore
Every magician knows there is no such thing as magic.
18:35 donsizemore
yes, there are new pem.* group_vars which, when populated, write out their respective keys and certs using servername for the filename
18:44 pdurbin
donsizemore: ok, maybe you can give me a sample main.yml with some self signed pem files in there or something. I only tried once but I assume I did something wrong.
18:44 donsizemore
@pdurbin i air-headedly suggested you try it without accommodating it in the code
18:45 donsizemore
see https://github.com/IQSS/dataverse-ansible/blob/master/defaults/main.yml#L10
18:45 donsizemore
you just place a bar after the each entry (say, cert: |)
18:45 donsizemore
then place your key or pem below, indenting with spaces to keep the yaml happy
18:46 donsizemore
if those group_vars are populated the role will write out the files and stick them in the ssl.conf template for you
18:47 pdurbin
oh, when I tried it there was no "pem:" :)
18:47 pdurbin
now I'm excited to try it again... after you merge :)
18:47 pdurbin
or after you resolve merge conflicts at least :)
18:47 pdurbin
next week
18:47 jri joined #dataverse
18:53 donsizemore
in the process of resolving them git merged them for me
18:53 donsizemore
even Travis is happy
18:54 donsizemore
i sent you a PR for dataverse-metrics as well
18:54 pdurbin
o.O
18:55 donsizemore
silly little shell script and cron example
18:55 donsizemore
heading for car dealership. back in a bit!
18:55 pdurbin
this looks great, will probably merge shortly, thanks!
19:07 pdurbin
Merged. New release of dataverse-metrics. Let Michel know: https://groups.google.com/d/msg/dataverse-community/fe7UZVJS18U/dYtu1b6AAgAJ
19:49 pdurbin
bjonnh: I haven't forgotten about you NMR thing. I printed it out and hope to read it basically now. Or next week if there's no time pressure. :)
19:52 bjonnh
sure
19:52 bjonnh
we will likely release a public draft early next week
19:52 bjonnh
but we can incorporate any change later
19:53 pdurbin
cool
20:07 donsizemore joined #dataverse
20:40 pdurbin
no time now. sorry. next week
20:40 pdurbin
have a great weekend, everyone!
20:40 pdurbin left #dataverse
22:36 bjonnh joined #dataverse
23:18 pameyer joined #dataverse
23:20 pameyer
bjonnh (if you're still on, or if you see the logs); I had a question about your NMR doc. do you have a ballpark estimate for how many spectra would be in a typical dataset?
