The IRC channel is #sakai on the freenode network, you can join using a web client at http://webchat.freenode.net/?channels=sakai and you can view the complete logs at http://sakai.iriscouch.com/irc/_design/viewer/index.html
13:39 <simong> Does anyone have any experience with converting schema.xml/solrconfig.xml from 3.1 to 4.0 ? 15:04 <stuartf> PhysX: sorry about the issues on SAKIII-6062, you should be safe to work on it though, I'm doing our big 2.8 upgrade today so probably won't get to look at OAE much 15:07 <PhysX> stuartf: no worries, there's one weird issue though 15:08 <PhysX> stuartf: for example the inserterbar has <a> tags for buttons (which we should change) that are now redirecting to a 404 15:08 <PhysX> stuartf: if we add a 'return false' they don't follow the link, but previously we didn't have to do that 15:08 <PhysX> stuartf: and I'm not sure why we suddenly need it, any ideas? 15:08 <stuartf> weird 15:09 <stuartf> nothing immediately comes to mind 15:10 <PhysX> stuartf: I have seen this before when we have JS errors, but nothing comes up in the console 15:14 <stuartf> well a lot of the quote changes and event binding changes were done with global search and replace, so it's possible that errors could creep in there 15:18 <PhysX> stuartf: does that mean that widgets who add html through javascript will have issues if they used double quotes in there? 15:19 <stuartf> PhysX: I don't think so, I think I caught all the places that had nested quotes 15:19 <stuartf> but it was a lot of lines, so it was tough to do really thorough review in the time allotted 15:21 <PhysX> makes sense it has 16k changes 15:44 <raydavis> simong, I have PLENTY of experience converting schema.xml : https://github.com/sakaiproject/nakamura/commit/26343d6311a93710d26316714967166e8ccd6ca6 15:44 <simong> Wow, that's quite the commit you've got there :) 15:45 <simong> I've actually figured it out though. The schema.xml that vivo uses seems to be compatible with 4.0.0. It was there search querystrings that weren't 15:45 <simong> But I'm now able to run vivo's indexes on the same solr server (in a seperate core though) and I can query it from Sakai 16:02 <nicola> anyone online right now? I have a question about the 1.4 release 16:04 <raydavis> simong, ah yes, the forward slash, right? 16:04 <raydavis> Very cool progress. 16:04 <PhysX> nicola: shoot 16:05 <simong> They were doing fieldName + ":" + query. Making it fieldName + ":\"" + query + "\"" fixed it 16:05 <nicola> when trying to build 1.4.0 mvn errored on javax.mail-activation - it couldn't find the proper version 16:08 <nicola> we checked out 1.4.0, looks like the main pom.xml says 1.4.1-SNAPSHOT 16:08 <nicola> and mail-activation doesn't seem to use the version that you specify its pom 16:08 <nicola> it constructs it from the parent version.....so its looking for the wrong version 16:08 <nicola> just grabbing the error now 16:10 <nicola> if we check out 1.4.0-release, THAT will build....but Anth did not mention that in his release email 16:10 <nicola> so im not sure what that is 16:10 <PhysX> nicola: what are you checking out now? 16:11 <Guest79018> I am bryan not Guest79018 16:11 <Guest79018> btw 16:11 <Guest79018> its version 1.4.0 from the release announcement 16:12 <Guest79018> build error: unable to find 16:12 <Guest79018> org.sakaiproject.nakamura:org.sakaiproject.nakamura.uxloader:jar:1.4.1-SNAPSHOT 16:12 <Guest79018> olp, different error 16:12 <Guest79018> hang on 16:13 <nicola> i think its the same root cause though......nakamura main pom says its version is 1.4.1-SNAPSHOT 16:14 <Guest79018> Do we need to use 3akai 1.4.1 instead of 1.4.0? 16:14 <PhysX> No, I don't think it should be looking for 1.4.1 16:15 <PhysX> Could you bring this on list? I'm not sure how much I can help you out with this. 16:15 <Guest79018> whats the list's email? 16:16 <nicola> yeah, which list? oae-dev@collab? 16:16 <PhysX> nicola, Guest79018 all info on http://collab.sakaiproject.org/mailman/listinfo/oae-dev 16:16 <PhysX> oae-dev@collab indeed 16:18 <simong> Could it be that you checked out the branch 1.4.0 rather than the tag org.sakaiproject.nakamura-1.4.0 ? 16:25 <nicola> ugh, i think we fixed our problem 16:25 <nicola> looks like nakamura 1.4.0 requires ux 1.4.1? 16:35 <MrVisser> ping froese 16:36 <froese> hey mrvisser 16:37 <MrVisser> froese: hey there! The /usr/local/sakaioae directory, I would like to chmod it to 755 instead of 750 so the driver machine can grab logs w/ scp.. any large problems with that? 16:37 <froese> nope. but change it in puppet too 16:37 <MrVisser> froese: absolutely. Thanks. 16:37 <froese> else its gonna come by and change it back 16:38 <MrVisser> froese: all changes I've made so far should be in puppet, so no worries there. 16:38 <froese> I'm gonna have notes for your PR this afternoon 16:38 <froese> Its a lot in one PR. Some of the subtree stuff needs to go into the individual modules with default arguments. 16:39 <MrVisser> froese: awesome. I didn't know git subtrees existed. will be a little tricky to work around. Some changes are one-off and so probably can't be pushed upstream unless we fork those modules. Tricky. 16:39 <froese> well the nice thing about subtrees is that you can have local modifications and still pull it upstream changes 16:40 <froese> they just rewrite the path to patch and present conflicts like a normal merge 18:29 <froese> Is there any documentation about the structure of sakai documents? 18:30 <froese> i trying to diagnose some content migration issues and I'm a little lost on the terminology 18:44 <duffy> raydavis, mrvisser: I have a question about convention and Solr queries 18:45 <duffy> raydavis, mrvisser: I've run into a user of LiteUserExistsServlet which is failing because the incoming userid is not escaped. My feeling is that query terms should be escaped as late as possible, and so I am looking to escape the userid within LiteUserExistsServlet. Does this gel with the way you guys have seen things in the search code? 18:53 <jonfelder> hey frose, are you around? 18:53 <raydavis> duffy, I agree in principle: escaping for a particular target (Solr/Lucene in this case) should happen in the code that knows the target. 18:54 <MrVisser> duffy: absolutely. 18:54 <raydavis> But AnythingServlet seems like too high a layer in that case -- shouldn't the escaping happen at the user-exists service instead? 18:54 <raydavis> Since user exists should use both Solr & Sparse services, one with escaping & one without. 18:55 <raydavis> (This is from memory, so I might have it wrong.) 18:55 <MrVisser> jonfelder: just fixing froese 's tag for you so he gets the ping ^^ 18:55 <jonfelder> woops 18:56 <froese> haha here i am jonfelder 18:56 <jonfelder> so I'm working on clustering 18:56 <jonfelder> I believe I followed the confluence page 18:56 <jonfelder> however when I logged into one node 18:56 <jonfelder> and when I log into a second node it acts as though it's the first time 18:57 <jonfelder> I suspect it has something to do with the local data store 18:57 <froese> are they both connected to the same solr server and database servers? 18:57 <jonfelder> yes 18:57 <froese> you have apache in front? 18:58 <jonfelder> yes 18:58 <froese> (just clarifying) so you're accessing both machines via the same URL, just turning one or the other app server off to test? 18:58 <jonfelder> no 18:58 <jonfelder> I'm accessing the individual node not through the balancer 18:59 <froese> i don't think that is going to work. When you hit the app server via its own hostname (or not he general service url) the cookie is going to be set for that domain 18:59 <froese> i.e. oae-app1.localdomain can't read cookies set for oae-app2.localdomain 18:59 <jonfelder> I see, so the cookie is what tells calcentral whether or not it is your first login? 19:00 <froese> its the login period. the cookie is your session. 19:00 <froese> In a typical cluster you'd have one apache serving a URL like oae.localdomain 19:00 <jonfelder> yeah I get that 19:00 <froese> The cookie is set per-domain 19:00 <duffy> raydavis, mrvisser: thanks :) 19:01 <jonfelder> but when I login to the second node I get the message for accepting my first login…I figured that info would be stored in data base 19:01 <froese> ohhhh i see what you're saying. 19:01 <froese> that I'm not sure of. 19:01 <froese> it should be stored as part of the user data in the DB 19:01 <froese> but i'm not sure where 19:01 <jonfelder> that's what I would think 19:01 <froese> open up the web developer console 19:01 <froese> you might see a POST request to update that information. 19:02 <jonfelder> would the jackrabbit store have anything to do with this? 19:03 <froese> nope. it would be in sparse and solr 19:03 <jonfelder> as a test I just shutoff apache on the first node 19:03 <froese> nothing gets written to jackrabbit except UI assets (templates, css, js) 19:04 <jonfelder> and connected to the balancer instead 19:04 <jonfelder> same result 19:04 <froese> use the web inspector or firebug to watch web requests in the background 19:04 <froese> when you click to accept /close that welcome message it should send data back to the server in a POST request 19:05 <raydavis> froese, minor correction: any resources that get loaded from the bundle get put in Jackrabbit. That also includes search templates & other Nakamura-hosted stuff, most of it starting at "/var". 19:06 <froese> thanks raydavis, all of that stuff is loaded from bundles and are determined at build time. Once they're loaded on startup there aren't really any writes to jackrabbit 19:07 <froese> unless you add templates or other assets by hand 19:07 <froese> but the app doesn't expose that 19:07 <jonfelder> ok i'm not overly familiar with firebug 19:07 <jonfelder> i've got it open 19:07 <jonfelder> it's showing me the html right now 19:08 <froese> the net panel is the one that shows network requests 19:08 <jonfelder> ok i'm on the console tab 19:09 <froese> not console, net 19:09 <jonfelder> I see the get requests…I'm going to login 19:09 <froese> network requests 19:09 <jonfelder> ok 19:09 <froese> ok once you do DONT click the welcome message 19:09 <froese> login, clear the net panel, then click the message 19:09 <jonfelder> ok so welcome message is up…going to clear, click the boxes, and accept correct? 19:11 <jonfelder> froese? 19:11 <froese> yes 19:11 <jonfelder> ok 19:12 <jonfelder> I have a few post requests 19:14 <jonfelder> what am I looking for? 19:14 <froese> I'm not sure honestly 19:14 <froese> one sec 19:15 <froese> on my local build of 1.5.0-SNAPSHOT it just faded away before I could click it 19:16 <jonfelder> i'm trying something else now 19:16 <jonfelder> I'm adding content to my library 19:16 <jonfelder> on node 2 19:16 <jonfelder> then i'm going to turn off apache on node 2 and connect with node 1 19:18 <jonfelder> file is not there 19:18 <jonfelder> I have no clue what's going on 19:19 <froese> the file is there or no record of the file at all? 19:20 <froese> sounds like app2 is not connected to solr 19:21 <jonfelder> well perhaps both aren't 19:21 <froese> wait, you have apache on each app server machine? 19:21 <jonfelder> yes 19:21 <jonfelder> apache is acting as a proxy on each machien 19:21 <froese> why? 19:21 <froese> you just need one apache on another machine that proxy's to 8080 and 8082 on the app servers 19:22 <jonfelder> i'm not really sure as to the reasoning 19:24 <jonfelder> is there a log which would indicate what the deal with solr is? 19:24 <froese> so you have a third apache parodying to the other 2? 19:24 <froese> go to the app server http://appserver/system/console 19:25 <froese> click configuration 19:25 <jonfelder> there is a balancer which sends the requests to the other two 19:25 <jonfelder> I do not know if the balancer runs apache. 19:25 <froese> ok 19:25 <froese> Look for the SolrServerServiceImpl on the configuration page 19:27 <froese> click the edit button and verify that the value in the field is "remote" 19:27 <jonfelder> great it says I don't have permission to access the console 19:27 <froese> (no quotes in the actually field) 19:27 <froese> quit your browser and log back in as admin 19:28 <jonfelder> no something else is up 19:28 <jonfelder> I get no option to login period 19:28 <jonfelder> when I go to system/console 19:28 <jonfelder> I just get permission denied 19:28 <froese> did you disallow access to that URL via apache? 19:28 <jonfelder> no 19:28 <jonfelder> i need to figure this out 19:28 <jonfelder> hang on 19:29 <froese> i have to get back to work 19:30 <jonfelder> ok 19:41 <jonfelder1> froese, when you get a chance…it's set to remote and the remoteurl is what it should be…and I can access that url from the node if I telnet to it 20:32 <froese> jonfelder1, thats a good sign 20:32 <froese> is there anything in sling/solr/nakamura? 21:20 <GitHub189> [nakamura] thecarlhall pushed 2 new commits to master: http://git.io/yGPrrg 22:57 <jonfelder1> hey froese are you there? 22:57 <froese> jonfelder1 yup 22:57 <froese> leaving in 10 though 22:57 <froese> whats up? 22:57 <jonfelder1> ok i'm making progress slowly but surely 22:57 <froese> awesome 22:57 <jonfelder1> i have another question 22:57 <froese> progress is good 22:58 <jonfelder1> if I start up node 1 22:58 <jonfelder1> and then start up node 2 with an empty repository I get issues 22:58 <jonfelder1> if I copy the entire sling directory from node 1 to node 2 it appears to work 22:59 <jonfelder1> my question is when I add node 3 22:59 <jonfelder1> do I need to stop nodes 1 and 2 22:59 <jonfelder1> and then copy the whole thing over? 22:59 <froese> what issues are you seeing? 22:59 <jonfelder1> complaints about the repository being created by another factory 22:59 <jonfelder1> I don't have the error in front of me 23:00 <froese> i can't do much without the errors 23:00 <froese> which repository? JCR or SparseMapContent? 23:00 <froese> you should't copy the whole sling directory over 23:00 <jonfelder1> ok what should I copy over? 23:00 <jonfelder1> just the configs? 23:00 <froese> are you sure the nodes are connecting to your database? 23:00 <froese> JUST sling/config/org/sakaiproject 23:01 <jonfelder1> ok 23:01 <froese> if you copy over sling/config/org/apache/ you're going to get jackrabi 23:01 <froese> jackrabbit errors 23:01 <jonfelder1> I'm going to mess around with it a bit more 23:01 <jonfelder1> I might catch you tomorrow with the errors 23:01 <jonfelder1> if I see them 23:02 <jonfelder1> one other thing, when I blow away the database do I have to do anything to solr? 23:02 <jonfelder1> like clear the indexes or anything? 23:02 <froese> run netstat -a -n on the app nodes and make sure they're connecting to your DB and solr 23:02 <jonfelder1> i'm actually pretty sure they are now 23:02 <froese> be certain 23:02 <jonfelder1> because I can add a file to the library on one node and it shows up on the other 23:03 <jonfelder1> I'm going to blow away the database because it's kind of wonky right now 23:03 <jonfelder1> and then restart both nodes 23:03 <jonfelder1> do I have to do anything to solr when I do that? 23:04 <froese> either stop solr, delete the index data, and restart 23:04 <jonfelder1> ok where is the index data housed? 23:04 <froese> OR go into the system/console on either of the nodes and go to the "Solr Index" tab and trigger a full re-index 23:04 <froese> wherever you defined as solr.home 23:10 <froese> ta run 23:11 <froese> later! _______________________________________________ oae-dev mailing list oae-dev@collab.sakaiproject.org http://collab.sakaiproject.org/mailman/listinfo/oae-dev
