IQSS logo

IRC log for #dataverse, 2016-08-01

Connect via chat.dataverse.org to discuss Dataverse (dataverse.org, an open source web application for sharing, citing, analyzing, and preserving research data) with users and developers.

| Channels | #dataverse index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

All times shown according to UTC.

Time S Nick Message
13:19 bsilverstein joined #dataverse
13:20 pdurbin bsilverstein: mornin'! Do you have enough to keep you busy?
13:20 bsilverstein pdurbin: for the time being yes!
13:21 bsilverstein will keep you posted
13:21 pdurbin awesome. I'm deploying the confirm email branch to https://shibtest.dataverse.org ... please just let me know whenever you want me to deploy the latest code from that branch
13:42 pameyer joined #dataverse
14:06 jeffspies______ joined #dataverse
14:06 skay_ joined #dataverse
14:08 donsizemore joined #dataverse
14:11 JonathanNeal joined #dataverse
14:17 LyndsySimon joined #dataverse
14:30 donsizemore joined #dataverse
14:32 donsizemore @pdurbin: knock knock?
14:35 pdurbin donsizemore: who's there?
14:35 donsizemore @pdurbin a no-re-indexing whinypants
14:35 pdurbin hmm
14:35 donsizemore @pdurbin couple of quick questions, if this shouldn't go to dataverse_support instead?
14:36 pdurbin I'll let you know :)
14:37 pdurbin if I don't know the answer :)
14:37 donsizemore @pdurbin: i see the "clear solr" curl command in the "fresh reinstall" section. check.
14:38 pdurbin donsizemore: you should probably look at this instead: http://guides.dataverse.org/en/4.4/installation/administration.html#solr-search-index
14:38 donsizemore @pdurbin: excellent. i found (somewhere) an SQL dvobject statement that wasn't working. thank you!
14:39 pdurbin hmm, not working is not good
14:39 donsizemore @pdurbin: question 3: is there a better way to trigger a re-index on 3.6 than through the web interface? everything comes back for 4.n
14:39 donsizemore @pdurbin or should the 4.n curl commands work under 3? same version of solr
14:40 donsizemore @pdurbin: I think where I was breaking down was in importing our production 3.6 machine to a scratch 3.6 machine... but not re-indexing THAT machine before proceeding to export DDIs
14:40 pdurbin reindexing in DVN 3.x was a pain, if I recall correctly. I don't know. The Dataverse 4 curl commands for reindexing do not work on DVN 3.x
14:40 donsizemore @pdurbin I can just do it through the web interface. many thanks!
14:41 pdurbin sure thing
15:07 pdurbin donsizemore: question for you about Shib
15:08 donsizemore @pdurbin: yes, sir. UNC is on Shib3 now, BTW =)
15:11 pdurbin donsizemore: the R&S form requires you to state that metadata is being refreshed automatically. Have you cracked this nut?
15:12 donsizemore @pdurbin UNC's Identity Management folks handle this for me at the SP level (at least I thought?)
15:13 pdurbin donsizemore: but you *are* refreshing metadata periodically?
15:13 donsizemore @pdurbin though I do specify this in my shibboleth.xml:
15:13 donsizemore <MetadataProvider type="XML"           url="http://md.incommon.org/InCommon/InCommon-metadata.xml"           backingFilePath="InCommon-metadata.xml" maxRefreshDelay="3600">
15:14 pdurbin donsizemore: ok. http://guides.dataverse.org/en/4.4/installation/shibboleth.html doesn't mention this. Do you feel like making a pull request to add it?
15:15 donsizemore @pdurbin I don't feel qualified to make a pull request of anyone, but I'm happy to try =)
15:15 pdurbin :)
18:46 andrewSC joined #dataverse
18:46 andrewSC hi all
18:46 andrewSC I'm trying to tune the performance of my install
18:47 andrewSC I'm noticing in particular, when I click the "+" button next to the Author (when creating a new dataset) that the request takes about 400ms
18:47 andrewSC is that "normal"?
19:04 pdurbin andrewSC: hi! I haven't tested that in particular...
19:05 pdurbin on my laptop it seems quicker than that. not slow at all
19:05 andrewSC pdurbin: same
19:05 pdurbin andrewSC: can you reproduce the slowness at https://demo.dataverse.org ?
19:06 andrewSC I'm walking through this guide though: http://blog.c2b2.co.uk/2013/07/glassfish-4-performance-tuning.html
19:06 andrewSC going to see if that makes much of a difference
19:06 andrewSC I'm on the glassfish part now
19:08 pdurbin that's a nice blog post
19:08 andrewSC pdurbin: yeah, it seems to take around 550-600ms to add a new Author field for a new dataset
19:10 pdurbin andrewSC: are you using a tool to time this?
19:10 andrewSC pdurbin: chrome dev console network requests
19:10 pdurbin ah, ok
19:32 andrewSC pdurbin: I've set those settings and don't notice any performance improvement... I do however get the same speeds as the demo site so I assume I have everything configured correctly..
19:33 andrewSC oh well
19:34 pameyer andrewSC - have you looked server side?
19:34 andrewSC pameyer: all my tweaks were server side?
19:35 pameyer sorry - I'd missed some of the chat history.  I'm just wondering if the slowdown is glassfish, postgres, or a query to the authentication provider
19:38 andrewSC pameyer: oh no worries, I think it's glassfish honestly, but I've tweaked it pretty good (per that perf guide) and i'm still seeing the same speeds at the demo site
19:38 andrewSC can i assume the demo site has been perf. tuned?
19:40 pameyer I don't have any info about the system configuration for the demo site.
19:41 pameyer It may not be directly related to your observation; but a few weeks back a co-worker noticed that list of files was slow (for intentionally very large numbers of files), and at least part of that was postgres
19:42 pameyer haven't investigated further
19:47 pdurbin andrewSC: I'm not sure if the demo site has been tuned much or not. support@dataverse.org would know

| Channels | #dataverse index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

Connect via chat.dataverse.org to discuss Dataverse (dataverse.org, an open source web application for sharing, citing, analyzing, and preserving research data) with users and developers.