Time
S
Nick
Message
00:49 jri joined #dataverse
03:49 jri joined #dataverse
06:28 juancorr joined #dataverse
06:50 jri joined #dataverse
11:54 donsizemore joined #dataverse
12:20 juancorr
jri: Hi!. You have a clean installation in http://oaimadrono.uned.es:8080 . The password for dataverseAdmin is opeanaire0
12:21 juancorr
Do you need a populate installation?
12:37 jri
Hello juancorr, I think your answer is not for me :/
13:49 pdurbin
juancorr: thanks! I'll let Julian know!
14:11 dataverse-user joined #dataverse
14:11 cthao joined #dataverse
14:18 cthao
For https://github.com/IQSS/dataverse/issues/4762 , would there be a need to change the java source code?
14:19 pdurbin
cthao: I don't believe so. I'm sorry for the hack I put in to run Solr out of /tmp. I was in a hurry. :(
14:20 juliangautier joined #dataverse
14:22 cthao
pdurbin: No problem :D So let's say we want to have replications of the index, the change would be in solrconfig.xml?
14:22 cthao
I think I
14:23 cthao
I think I'm oversimplifying things but I don't exactly know how Solr works in depth
14:23 pdurbin
cthao: we might need to ask on the Solr mailing list. I'm subscribed and can help back up anything you ask there.
14:24 pdurbin
Honestly, I would focus first on using an OpenShift-compatible Solr image already on Docker Hub and help us get away from creating our own.
14:25 pdurbin
And *then* get fancy with replication and such. Does that make sense?
14:25 juliangautier
Hi Juan! Thanks for setting up that installation! (I can populate it.)
14:26 cthao
pdurbin: I understand that. Dan was saying that the image might not fit in the context of StatefulSets.
14:27 cthao
I think the problem is the solr image might not be scalable since solr wants people to use their SolrCloud
14:27 pdurbin
cthao: you should definitely follow Dan's lead on this but you can mention to him what I said about Docker Hub if you want.
14:28 cthao
We checked that one out a couple days ago, but yes I will mention it again
14:28 cthao
there's some cool stuff here too
14:28 cthao
https://www.youtube.com/watch?v=PC8mYweMgV4&t=434s
14:33 pdurbin
Nice. Thanks. I added a link to the issue.
14:33 cthao
*thumbs up emoji*
14:33 pameyer joined #dataverse
14:34 pameyer
cthao , pdurbin: I'm still vaguely suspicious that something might break if postgres and solr aren't fully in sync
14:35 pdurbin
pameyer: I just think of Solr as a step behind Postgres, the source of truth.
14:35 cthao
Are you referring to Solr and Postgres for distributed use?
14:37 pameyer
I'm thinking that if postgres has a master for writes, which get replicated to secondary instances for reads, then there's the potential for glassfish to read from a a secondary postgres with older information that the solr instance
14:38 pameyer
I know that solr being down leads to dataverse failures, but I don't know how it fails if there are differences between database and search index
14:38 pdurbin
pameyer: but glassfish won't read from a non-master postgres until someone works on https://github.com/IQSS/dataverse/issues/4765
14:39 pameyer
right - it's not a problem that would show up without postgres scaling
14:40 pdurbin
Yeah. Basically no one should be setting replicas for postgres to be greater than one until 4765 has been resolved.
14:40 pdurbin
cthao: do you happen to know what Tommy is working on these days?
14:40 cthao
I do
14:40 cthao
I think the idea is to have a Secondary datasource , not a secondary db
14:41 pdurbin
cthao: what's he working on? :)
14:44 cthao
he's working on #4765, and I think he's using a design to achieve it. Could I have your email so I send you the presentation we gave last friday?
14:45 pdurbin
cthao: sure it's on https://www.iq.harvard.edu/people/philip-durbin
14:46 cthao
sent
14:47 pdurbin
cthao: got it. Thanks. Can I add it to the issue?
14:47 cthao
sure
14:47 pdurbin
cthao: thanks. By the way... s/DataVerse/Dataverse/ (slide 4)
14:48 pdurbin
Man, I bet this was an interesting talk. All Dataverse-y. How did it go?
14:48 cthao
It was very well received
14:49 cthao
there was a small discussion with Orran Krieger at the end, but the general vibe was positive
14:49 juliangautier left #dataverse
14:51 pdurbin
Nice.
14:54 pameyer joined #dataverse
14:55 pdurbin
Ok, slides added. Thank again! pameyer you might want to take a look: https://github.com/IQSS/dataverse/issues/4765
15:01 pameyer
sometimes slides without the presentation aren't that clear :( of course, I'm guilty of that too
15:03 pdurbin
cthao: was it recorded?
15:05 cthao joined #dataverse
15:06 cthao
it was
15:06 pdurbin
cthao: oh! Can you please add a link to the issue?
15:16 cthao
I should double check with Red hat before hand
15:16 cthao
I don't think it's public
15:32 donsizemore joined #dataverse
15:37 pameyer joined #dataverse
15:56 pameyer joined #dataverse
15:58 pameyer joined #dataverse
16:25 jri joined #dataverse
16:26 pameyer joined #dataverse
16:29 pameyer
cthao: might not be worth the trouble to get it made public. I was slightly confused about which parts of the diagrams corresponded to what, but those will get cleared up when code comes in (or other ways of talking about it)
16:30 pameyer
@donsizemore - fyi https://docs.google.com/document/d/1pMo3Pb13setm99NdQCGzk-EpMywJtnNyGTb3ZMH39KI had some interested in dataverse-ansible
17:15 donsizemore joined #dataverse
17:17 donsizemore
@pameyer if i had my official Twilight Zone Stopwatch (tm) and could get done all the things I want to get done, dataverse-ansible would get a fair amount of retooling
17:22 pdurbin
donsizemore: like Pete is saying, Jamie was on the call and said she's making progress
17:29 pameyer joined #dataverse
17:31 pameyer
@donsizemore: if you find a reliable supplier for working Twilight Zone Stopwatch (tm) 'es, please let me know - I could use a few
17:32 pdurbin
time enough at last
17:41 juancorr joined #dataverse
19:01 jri joined #dataverse
19:06 pameyer joined #dataverse
19:10 cthao joined #dataverse
19:11 cthao
pdurbin: it might be worth asking Tommy directly. I am not exactly clear on his part of the work.
19:12 cthao
anyway, are you aware of SolrCloud?
19:12 cthao
I think it's a distributed version of Solr
19:13 pameyer
I'm vaguely aware of SolrCloud as a clustered solr setup (aka - `solr create_cloud` vs `solr create_core`
19:13 pameyer
but only vaguely; pdurbin probably knows more then me
19:24 cthao joined #dataverse
19:25 cthao joined #dataverse
19:48 pdurbin
I've heard of it.
20:28 pameyer joined #dataverse
21:05 pameyer joined #dataverse
21:14 juliangautier joined #dataverse
21:30 pameyer joined #dataverse
21:57 pameyer joined #dataverse
23:02 jri joined #dataverse
