Time
S
Nick
Message
08:26 jri joined #dataverse
08:27 juancorr joined #dataverse
08:40 jri_ joined #dataverse
12:27 donsizemore joined #dataverse
12:40 * pdurbin
hopes donsizemore noticed his Twilight Zone reference yesterday
12:44 donsizemore
I did, I did! I was also going to fix that guy's sudo dealy, but had a puppy dog in my lap with morning coffee
13:09 jri joined #dataverse
13:14 donsizemore
@pdurbin I'm going to add the "destroy" verbiage to the Admin section of the Native API page, then link to it from my solr PR, sound good?
13:29 donsizemore
@pdurbin I ask because the endpoint isn't beneath /admin by URL but does require the API token belong to a superuser
13:36 donsizemore
@pdurbin and, now that i review my own notes, i didn't have to pass an API token(!)
13:38 donsizemore
@pdurbin wait. yes i did. whew.
13:52 pdurbin
free as in puppy
13:52 pdurbin
donsizemore: one pull request for multiple issues is just fine. Thanks!
13:53 jri_ joined #dataverse
14:16 pameyer joined #dataverse
14:32 pdurbin
pameyer: mornin'. Are you running PostgreSQL 8.x or 9.x?
14:33 pameyer
pdurbin: I'll take a look, but relatively sure something in the 9.x series
14:33 donsizemore joined #dataverse
14:34 pdurbin
Good. I just opened https://github.com/IQSS/dataverse/issues/5649
14:34 donsizemore
@pdurbin UNC Dataverse is on 9.3
14:34 pdurbin
I thought I remembered you using pg8-createsequence-prep.sql but I could be wrong.
14:35 donsizemore
(and hoping to migrate to 9.6 once our new digs in Research Computing are ready)
14:35 pameyer
if I'm remembering correctly, the createsequence bits were done when pg 8 was the dataverse recommended version
14:35 pdurbin
sounds right
14:35 pameyer
I'd been under the impression that there was a recommended postgres version; but I vaguely remember fiddling w\ docker-aio to make sure it was using that version from the postgres repo (vs the cent repo)
14:36 pdurbin
donsizemore: good plan. That way you'll be in sync with Harvard Dataverse, which upgraded to PostgreSQL 9.6 as part of https://github.com/IQSS/dataverse/issues/4709
14:37 pdurbin
donsizemore: you might remember helping us upgrade the drivers. Thanks again. :)
14:37 pdurbin
pameyer: ah, let me look at Vagrant and see if I need to add it to the definition of done.
14:38 pdurbin
# switching to postgresql-9.6 per #4709
14:38 pdurbin
phew
14:38 pdurbin
probably thanks to one of you two
14:38 pdurbin
56043ac348 (donsizemore
14:38 pdurbin
thanks, Don
14:40 pdurbin
What's the default version of postgres on centos 7?
14:41 pameyer
looks to me like 9.2
14:41 pdurbin
Ouch. Not new enough for free Flyway.
14:41 pameyer
not sure if that'll vary with minor versions of cent7; but wouldn't be surprised if it did
14:42 pdurbin
me too
14:42 pameyer
welp, I asked about that a couple times
14:42 pdurbin
you sure did
14:42 pdurbin
does docker-aio need to be updated?
14:43 pameyer
9.6 in there
14:44 pdurbin
phew
14:44 pdurbin
our openshift.json says centos-postgresql-94-centos7
14:45 pdurbin
nice, an upgrade from 9.1 to 9.3 yesterday (!) at https://github.com/IQSS/dataverse-docker/commit/8c912ac
14:48 pdurbin
Opinions on the version used in dataverse-kubernetes? I don't want to bias anyone but from a glance at this I'm thinking postgres 9.2 (centos 7 default): https://github.com/IQSS/dataverse-kubernetes/search?q=postgresql&unscoped_q=postgresql
14:53 pameyer
I'd guess folks will use whatever's in the repos by default, and may or may not read the docs that say "use this non-default version"
14:53 pdurbin
Sure. That's what I was doing on the phoenix server. I'm lazy. I'm checking my dev1 server now. Had to reboot it, the poor thing.
14:54 pdurbin
Both phoenix and dev1 are CentOS 6. T-T
14:58 pameyer
I don't know how far back the postgres repo goes for cent support - they might have a 9.x one that'll run on cent6
14:59 donsizemore
@pdurbin Flyway dropped it because 9.3 is EOL and no longer maintained upstream. Red Hat will continue to backport security patches into 9.2 so it isn't the end of the world, but...
15:04 pdurbin
donsizemore: sorry, are you saying 9.2 is EOL or 9.3 is EOL?
15:13 donsizemore
both, as far as the postgres maintainers are concerned. Red Hat backports security patches into their distributed RPMs, so CentOS gets them as well
15:14 pdurbin
Ok. Interesting that Flyway supports postgres 9.3 for free if it's EOL. Thanks.
15:14 pdurbin
donsizemore: and sorry about pull request confusion. I see now that you went for separate pull requests. In QA now.
15:16 donsizemore
@pdurbin quick question about metadata blocks?
15:17 donsizemore
thumai sent me an APA block she wants to use. she made the TSV. I upload it and can see it in the GUI . now I must edit the solr schema.
15:19 donsizemore
I have my list of copyFields and field names. i now wedge in the datasetfields into solr. copyfields are controlled vocabulary?
15:35 pdurbin
donsizemore: I wandered off to standup. Every field gets a copyfield.
15:41 pdurbin
you ran curl http://localhost:8080/api/admin/index/solr/schema I assume
15:56 pdurbin
donsizemore: here's where I describe copyField as the way content gets copied to a "catchall" field called "_text_" which is what powers basic search: https://github.com/IQSS/dataverse/blob/v4.11/conf/solr/7.3.1/schema.xml#L37-L38
16:21 pdurbin
Oh no! https://blog.waffle.io/farewell-from-waffle-%EF%B8%8F-794da4a72851
16:45 pdurbin
created an issue: https://github.com/IQSS/dataverse/issues/5651 ... so sad T-T
17:01 pameyer
this might be something glitching on my end, but it looks like dataverseAdmin is set as an admin twice (from the roleassignment table)
18:41 pdurbin
pameyer: not sure, but please feel free to leave a comment at https://github.com/IQSS/dataverse/issues/5652
18:41 pdurbin
donsizemore: still there?
18:56 donsizemore
@pdurbin yes sir! i've upgraded a test instance holding old live data to 4.9.4 to try again with my solr entries
18:57 donsizemore
@pdurbin i didn't know my ansible installation instructions were behold to the installation docs, sorry.
18:58 pdurbin
donsizemore: heh. They aren't! But thanks for merging. :) I was going to ask about your "destroy" pull request.
18:59 donsizemore
i'll fix that.
18:59 pdurbin
It's something else.
18:59 donsizemore
whut i do
18:59 pdurbin
I was wondering if you are willing to make the examples into copy and pasteable full curl commands.
19:00 pdurbin
Mostly for people who don't know you have to add -X to get DELETE to work from curl. Does that make sense?
19:01 donsizemore
i like that other pages used the full curl command but kept my additions short to match the rest of the page.
19:01 donsizemore
i'm happy to make them complete
19:04 pdurbin
please make them complete
19:04 pdurbin
I'm trying to slowly stamp out the non complete ones.
19:04 donsizemore
and i apologize for being dense, but to add thu-mai's stuff to schema.xml... i'm adding a "field name=" and a "copyfield=" entry for each? or just the copyfield
19:05 donsizemore
i'll stamp them all out in that page, i like complete examples much better
19:05 pdurbin
ok, but maybe in a separate issue/pull request?
19:06 pdurbin
you have to have "field name". Otherwise Dataverse will thrown exceptions when fields it wants aren't there
19:07 pdurbin
I guess copyfield is optional but if you don't add copyfield your custom fields won't be available via basic search.
19:07 donsizemore
excellent. will do
19:07 pdurbin
"foo" is basic search
19:07 pdurbin
"title:foo" is advanced search
19:07 pdurbin
or "authorName:foo"
19:07 pdurbin
with a :
19:07 pdurbin
so you probably want the copyfield lines
19:08 pdurbin
my mental model for this is that there's a bucket called _text_ that basic search uses. you decide what goes in that bucket
19:08 pdurbin
and you use copyfield to put stuff in that bucket
19:08 donsizemore
got it
19:09 donsizemore
(or at least, i understand it for the next 10 minutes)
19:09 pdurbin
:)
19:10 pdurbin
who knows if I have it right
19:14 donsizemore
well you do, of course!
19:14 pdurbin
Mr Solr
19:15 donsizemore
mr. solr should be in cancun this week
19:16 donsizemore
the kids are just starting to come back early from spring break and i'm insanely jealous
19:28 pdurbin
tanned and happy, I bet
19:31 pdurbin
https://twitter.com/search?q=%23waffleshutdown
20:59 donsizemore joined #dataverse
21:28 pdurbin
now I'm scrolling through q=waffleio
21:28 pdurbin
so sad
21:30 pdurbin
whelp, have a good weekend, everyone
21:30 pdurbin left #dataverse
23:39 pameyer
pdurbin: thanks for the link (sorry about the lag). 5652 looks possible different to me; but not something I'm going to sort out tonight
