uPortal IRC Logs-2011-04-18

Owned by Apereo Infrastructure
Last updated: Apr 18, 2011Version comment
10 min read

[10:15:57 CDT(-0500)] <mklein21005> when i try to login it displays a uPortal error that says an error is preventing it from rendering. When i hit refresh it successfully log in
[10:16:20 CDT(-0500)] <mklein21005> this is the error information it give me http://dpaste.com/533234/
[10:16:33 CDT(-0500)] <mklein21005> any idea why this could be happening?
[10:17:04 CDT(-0500)] <EricDalquist> what versioon of uportal?
[10:17:59 CDT(-0500)] <mklein21005> 3.2.4
[10:19:06 CDT(-0500)] <EricDalquist> and you get this everytime you login with this user?
[10:19:09 CDT(-0500)] <EricDalquist> or with all users?
[10:19:43 CDT(-0500)] <mklein21005> just users that authenicate against AD
[10:20:07 CDT(-0500)] <mklein21005> the local users still work like "admin"
[10:20:12 CDT(-0500)] <EricDalquist> do you have DLM layouts setup ?
[10:21:36 CDT(-0500)] <mklein21005> ummm im not sure
[10:21:49 CDT(-0500)] <mklein21005> how could i check
[10:21:51 CDT(-0500)] <mklein21005> that
[10:22:07 CDT(-0500)] <EricDalquist> did you customize uportal-impl/src/main/resource/properties/dlm.xml
[10:22:21 CDT(-0500)] <EricDalquist> I'm wondering if there is a DLM fragment that only applies to users logging in via AD that has some problem with it
[10:23:03 CDT(-0500)] <mklein21005> nooo i haven't touched that file
[10:23:55 CDT(-0500)] <EricDalquist> and it only happens the first time an AD user logs in?
[10:24:01 CDT(-0500)] <EricDalquist> or every time that user logs in?
[10:24:16 CDT(-0500)] <mklein21005> it happens everytime
[10:25:32 CDT(-0500)] <mklein21005> it started doing this after i tried to enable clearview
[10:25:48 CDT(-0500)] <EricDalquist> CAS ClearPass?
[10:25:50 CDT(-0500)] <mklein21005> but then i decided against it so i undid my changes
[10:25:57 CDT(-0500)] <mklein21005> yeah
[10:26:01 CDT(-0500)] <mklein21005> thats what i ment
[10:26:08 CDT(-0500)] <EricDalquist> do you have a link to the instructions for enabling that?
[10:26:35 CDT(-0500)] <mklein21005> https://wiki.jasig.org/display/UPM32/ClearPass+Integration+-+Maven+Overlay
[10:27:19 CDT(-0500)] <EricDalquist> thnx
[10:27:25 CDT(-0500)] <EricDalquist> I'll be back in ~ 15 minutes
[10:27:30 CDT(-0500)] <mklein21005> ok
[10:51:31 CDT(-0500)] <athena> EricDalquist: when you're around again, i'm having some strange issues w/ the notice plugin
[11:05:14 CDT(-0500)] <EricDalquist> I'm back but a bit swamped
[11:05:24 CDT(-0500)] <EricDalquist> our HR system switch is not going so smooth
[11:06:15 CDT(-0500)] <EricDalquist> what are the issues with the notice plugin athena/
[11:06:43 CDT(-0500)] <athena> it doesn't seem to be finding a couple licenses that are in the license file
[11:06:55 CDT(-0500)] <athena> for example: [WARNING] Failed to find license info for: xalan:serializer:jar:2.7.0:compile
[11:07:09 CDT(-0500)] <athena> but it looks to me like that artifact is already in the file in svn?
[11:16:53 CDT(-0500)] <EricDalquist> huh
[11:17:28 CDT(-0500)] <EricDalquist> yeah I see it there
[11:17:30 CDT(-0500)] <athena> yeah i can't really figure out what's going on
[11:17:34 CDT(-0500)] <EricDalquist> that is really weird
[11:17:41 CDT(-0500)] <athena> there are two other dependencies it's doing that for as well
[11:17:48 CDT(-0500)] <EricDalquist> any extra info if you run maven in debug mode?
[11:18:06 CDT(-0500)] <athena> let me try
[11:21:12 CDT(-0500)] <athena> http://pastebin.com/syvhXQ06
[11:21:54 CDT(-0500)] <EricDalquist> [DEBUG] Found org.jasig.maven.notice.lookup.ArtifactLicense@22cb4138[groupId=xalan, artifactId=xalan, version={}, name=<null>, license=Apache License, Version 2.0] with match null for: xalan:xalan:2.7.0
[11:21:55 CDT(-0500)] <EricDalquist> hrm
[11:22:01 CDT(-0500)] <EricDalquist> that line is interesting
[11:22:45 CDT(-0500)] <EricDalquist> wait
[11:22:51 CDT(-0500)] <EricDalquist> no that's not the same artifact
[11:23:18 CDT(-0500)] <athena> yeah
[11:23:37 CDT(-0500)] <athena> hang on
[11:24:06 CDT(-0500)] <athena> http://pastebin.com/DVwPQFVg
[11:24:08 CDT(-0500)] <athena> that's better
[11:24:33 CDT(-0500)] <athena> don't really understand what's going on
[11:24:37 CDT(-0500)] <EricDalquist> wow
[11:24:43 CDT(-0500)] <EricDalquist> either do I
[11:25:08 CDT(-0500)] <EricDalquist> any indications the notice plugin is failing to load the license file?
[11:26:13 CDT(-0500)] <athena> no, i do see this line higher up:
[11:26:15 CDT(-0500)] <athena> [DEBUG] Mapped org.jasig.maven.notice.lookup.ArtifactLicense@648bfdea[groupId=xalan, artifactId=serializer, version={}, name=<null>, license=Apache License, Version 2.0] from: https://source.jasig.org/licenses/license-mappings.xml
[11:26:38 CDT(-0500)] <EricDalquist> yup
[11:26:43 CDT(-0500)] <EricDalquist> that is what is weird about that log
[11:26:48 CDT(-0500)] <EricDalquist> it looks like it finds the mapping
[11:26:51 CDT(-0500)] <EricDalquist> but then fails later on
[11:26:54 CDT(-0500)] <athena> yeah
[11:27:07 CDT(-0500)] <EricDalquist> do you have any maven proxies setup?
[11:27:07 CDT(-0500)] <athena> i don't get it
[11:27:11 CDT(-0500)] <athena> no
[11:27:35 CDT(-0500)] <athena> wonder if you'd get the same error
[11:27:52 CDT(-0500)] <EricDalquist> could you, just for humor, try renaming .m2/repository
[11:27:57 CDT(-0500)] <EricDalquist> and run it with a clean local repo
[11:28:11 CDT(-0500)] <EricDalquist> I do see the project connecting to a bunch of non-central repos in that log
[11:28:17 CDT(-0500)] <EricDalquist> maybe that's causing some sort of problem
[11:29:16 CDT(-0500)] <athena> i think it's still going to connect to those same repos - they're in the calendar portlet pom
[11:29:37 CDT(-0500)] <EricDalquist> :/
[11:30:00 CDT(-0500)] <EricDalquist> I really wonder if one of those has some bad data for xalan:serializer in it
[11:30:12 CDT(-0500)] <athena> guess it's possible
[11:30:48 CDT(-0500)] <athena> not sure what we would be able to do about it
[11:33:53 CDT(-0500)] <athena> you know, i'm not sure those artifacts exist in central
[11:34:03 CDT(-0500)] <athena> http://repo1.maven.org/maven2/javax/script/script-api/ returns a 404
[11:34:25 CDT(-0500)] <athena> so maybe that's the issue
[11:34:27 CDT(-0500)] <EricDalquist> right, that's the problem when not everything is in central
[11:34:44 CDT(-0500)] <EricDalquist> do doing a build and notice check with a clean local repo might be a good test
[11:35:08 CDT(-0500)] <mklein21005> @EricDalquist did you happen to notice anything in the instructions I sent you that might be causing my issues?
[11:36:00 CDT(-0500)] <EricDalquist> I haven't had a chance to look at the code yet
[11:36:06 CDT(-0500)] <EricDalquist> nothing obvious though (sad)
[11:43:13 CDT(-0500)] <athena> ok, it does seem to be because of missing dependencies
[11:43:22 CDT(-0500)] <athena> updated xalan:serializer to 2.7.1
[11:43:25 CDT(-0500)] <athena> which fixed that problem
[11:44:59 CDT(-0500)] <athena> script-api and js-engine are actually being included by cernunnos
[11:45:07 CDT(-0500)] <athena> not really sure where those dependencies disappeared to
[11:49:31 CDT(-0500)] <athena> ah.
[11:49:37 CDT(-0500)] <athena> adding the jasig 3rd party repo solved it
[11:49:44 CDT(-0500)] <EricDalquist> (smile)
[11:49:59 CDT(-0500)] <athena> didn't realize we still needed it (sad)
[11:50:19 CDT(-0500)] <EricDalquist> yeah crn uses a bunch of libs that aren't in central
[11:50:42 CDT(-0500)] <athena> yeah
[11:51:55 CDT(-0500)] <athena> least i'll be able to cut a release now!
[11:52:28 CDT(-0500)] <EricDalquist> (smile)
[12:02:32 CDT(-0500)] <rickard> Ran into this today: http://pastebin.com/G8nw73mA
[12:02:49 CDT(-0500)] <rickard> Followed by: http://pastebin.com/3QJkrQCH
[12:02:51 CDT(-0500)] <EricDalquist> rickard: try adding a validation query to your dbcp config
[12:03:00 CDT(-0500)] <rickard> Yeah, I just did. Hope it helps.
[12:03:10 CDT(-0500)] <rickard> I was just curious if you had run into it before.
[12:03:26 CDT(-0500)] <EricDalquist> it was caused by your jdbc driver and/or db
[12:03:41 CDT(-0500)] <EricDalquist> somehow the connection between the two dropped or was stale or something
[12:03:53 CDT(-0500)] <rickard> I added:
[12:03:54 CDT(-0500)] <rickard> <property name="testOnBorrow" value="true"/>
[12:03:55 CDT(-0500)] <rickard> <property name="validationQuery" value="SELECT 1"/>
[12:04:12 CDT(-0500)] <EricDalquist> sounds good ... if select 1 works for your db
[12:04:17 CDT(-0500)] <rickard> It does.
[12:04:38 CDT(-0500)] <rickard> Speaking of which, is it possible to use the uPortal DB pool from within my own portlets?
[12:04:51 CDT(-0500)] <rickard> Or do I have to setup connection pooling manually for the portlets?
[12:04:58 CDT(-0500)] <EricDalquist> http://pastebin.com/p0g00z7G
[12:05:02 CDT(-0500)] <EricDalquist> that is our standard DB config
[12:05:05 CDT(-0500)] <EricDalquist> nope
[12:05:10 CDT(-0500)] <EricDalquist> you need to do one pool for each portlet
[12:05:11 CDT(-0500)] <EricDalquist> OR
[12:05:16 CDT(-0500)] <EricDalquist> setup the pools in tomcat
[12:05:26 CDT(-0500)] <EricDalquist> and use JNDI in the portal and the portlets to access the pool
[12:05:39 CDT(-0500)] <EricDalquist> for robustness we do 1 pool for each portlet app
[12:05:50 CDT(-0500)] <EricDalquist> then if 1 portlet app goes bad it can't take out anything other than itself
[12:06:21 CDT(-0500)] <rickard> Ok.
[12:06:38 CDT(-0500)] <rickard> That's not how the standard db config looks in 3.2.4.
[12:06:47 CDT(-0500)] <rickard> If we're speaking of uportal-impl/src/main/resources/properties/contexts/datasourceContext.xml
[12:08:51 CDT(-0500)] <EricDalquist> we as in UW Madison
[12:08:57 CDT(-0500)] <rickard> Ah, ok (smile)
[12:09:00 CDT(-0500)] <EricDalquist> the standard db config in uportal needs some tweaks
[12:09:17 CDT(-0500)] <EricDalquist> also we can't include validation out of the box because there is no such thing as a universal validation query
[12:09:28 CDT(-0500)] <rickard> True
[12:09:42 CDT(-0500)] <rickard> Might be something to add to the wiki though.
[12:09:49 CDT(-0500)] <rickard> For each db type.
[12:09:59 CDT(-0500)] <EricDalquist> yeah
[12:10:08 CDT(-0500)] <EricDalquist> steve swinsburg had brought that up as welkl
[12:10:28 CDT(-0500)] <EricDalquist> I thought it would be need to have someone write a little code to do it automatically
[12:10:46 CDT(-0500)] <EricDalquist> put together a mapping of hibernate dilects -> validation queries
[12:10:51 CDT(-0500)] <EricDalquist> since we know the dialiect
[12:10:54 CDT(-0500)] <rickard> True
[12:11:54 CDT(-0500)] <rickard> I'm off. Thanks.
[12:20:22 CDT(-0500)] <EricDalquist> watching our problems here with this hrs switch I really wonder if it would be valuable for uportal to have some sort of QoS for portlet execution
[12:20:46 CDT(-0500)] <EricDalquist> like say no more than x% of the total rendering pool can be used by a single portlet definition
[12:20:59 CDT(-0500)] <EricDalquist> so that one poorly behaved portlet can't monopolize all the rendering threads
[12:22:08 CDT(-0500)] <athena> sounds ncie to me!
[12:22:21 CDT(-0500)] <EricDalquist> we're having issues where this one portlet is really slow
[12:22:28 CDT(-0500)] <EricDalquist> like 30s/request and always timesout
[12:22:43 CDT(-0500)] <EricDalquist> the problem then is that one portlet ends up consuming every rendering rhead
[12:49:12 CDT(-0500)] <athena> yeah
[12:49:15 CDT(-0500)] <athena> i've seen that happen
[12:52:39 CDT(-0500)] <athena> hrm.
[12:52:52 CDT(-0500)] <athena> is the fragment admin stuff working for you right now?
[12:53:13 CDT(-0500)] <EricDalquist> in trunk ... not sure
[12:53:18 CDT(-0500)] <EricDalquist> I can check maybe later today
[12:53:21 CDT(-0500)] <athena> ok
[12:53:22 CDT(-0500)] <athena> hmm
[12:53:27 CDT(-0500)] <EricDalquist> bit crazy right now
[12:53:29 CDT(-0500)] <EricDalquist> (sad)
[12:53:30 CDT(-0500)] <athena> yeah
[12:53:35 CDT(-0500)] <athena> will think of a way to get around it for now
[12:54:36 CDT(-0500)] <EricDalquist> seeing errors?
[12:54:42 CDT(-0500)] <athena> yep
[12:54:49 CDT(-0500)] <EricDalquist> paste a stack
[12:54:53 CDT(-0500)] <EricDalquist> I at least have time to look at that
[12:55:47 CDT(-0500)] <athena> http://pastebin.com/2zAVrUxx
[12:55:57 CDT(-0500)] <athena> i'll try using the calendar portlets import/export instead
[13:04:12 CDT(-0500)] <EricDalquist> duh, dumb error athena
[13:09:36 CDT(-0500)] <EricDalquist> I'll have a fix shortly
[13:13:54 CDT(-0500)] <EricDalquist> ok athena
[13:13:55 CDT(-0500)] <EricDalquist> should be fixed
[13:14:39 CDT(-0500)] <athena> thanks so much! (smile)
[13:16:21 CDT(-0500)] <EricDalquist> yeah
[13:16:25 CDT(-0500)] <EricDalquist> just forgot a null check
[13:16:39 CDT(-0500)] <EricDalquist> and it was never null ... unless you're logged in as afragment owner (tongue)
[13:24:55 CDT(-0500)] <athena> oh (smile)
[13:24:59 CDT(-0500)] <athena> thanks for fixing it!
[13:25:04 CDT(-0500)] <athena> will make things a bit easier
[13:40:49 CDT(-0500)] <athena> my eclipse instance is really misbehaving today
[14:02:31 CDT(-0500)] <mklein21105> @EricDalquist i found some errors in my cas.log. http://dpaste.com/533311/ for some reason it thinks my credentials are incorect
[14:09:17 CDT(-0500)] <athena> are you using localhost:8080 still?
[14:09:51 CDT(-0500)] <athena> EricDalquist: looks like user impersonation needs a bit of TLC as well: http://pastebin.com/2C7NpzU6
[14:10:04 CDT(-0500)] <EricDalquist> hrm, ok
[14:10:18 CDT(-0500)] <EricDalquist> does that have a Caused By?
[14:10:32 CDT(-0500)] <EricDalquist> or is that the root of the trace
[14:10:43 CDT(-0500)] <athena> that's it, unfortunately
[14:11:16 CDT(-0500)] <EricDalquist> ok
[14:11:27 CDT(-0500)] <EricDalquist> what user did you swap to?
[14:11:35 CDT(-0500)] <athena> guest
[14:11:38 CDT(-0500)] <mklein21105> yeah im still using localhost
[14:11:43 CDT(-0500)] <EricDalquist> and is this just on the initial render?
[14:11:46 CDT(-0500)] <athena> which i suppose is kind of a special case
[14:11:47 CDT(-0500)] <athena> yep
[14:11:59 CDT(-0500)] <mklein21105> im sort of setting up a demo for the school im working fo
[14:12:00 CDT(-0500)] <mklein21105> r
[14:12:28 CDT(-0500)] <athena> mklein21105: not quite sure what's going on, but it looks a bit like maybe the user is authenticated, but then the portal's not able to get a proxy ticket back?
[14:12:38 CDT(-0500)] <athena> is this the result of setting up clearpass?
[14:13:29 CDT(-0500)] <mklein21105> i was in the process of setting up clearpass but then changed my code back to how it was before
[14:14:19 CDT(-0500)] <athena> ah - sounds like something might be in between states
[14:17:00 CDT(-0500)] <mklein21105> what do you mean by in between states
[14:17:01 CDT(-0500)] <mklein21105> ?
[14:17:31 CDT(-0500)] <athena> kind of sounds like something didn't get completely changed back - like perhaps you're at a state that's in between clearpass and no-clearpass
[14:33:20 CDT(-0500)] <athena> argh
[14:33:27 CDT(-0500)] <athena> now i'm having trouble w/ the license plugin
[14:33:31 CDT(-0500)] <EricDalquist> (sad)
[14:33:35 CDT(-0500)] <athena> if i run "mvn license:check" all returns happiness
[14:33:38 CDT(-0500)] <EricDalquist> this is an ARG sort of day
[14:33:42 CDT(-0500)] <athena> no kidding
[14:33:51 CDT(-0500)] <athena> but if i try mvn release:prepare, returns errors
[14:33:58 CDT(-0500)] <EricDalquist> (tongue)
[14:34:06 CDT(-0500)] <athena> really don't know what to do about that one
[14:45:25 CDT(-0500)] <mklein21105> i've checked over my configurations and am pretty sure all my changes are correct
[14:54:17 CDT(-0500)] <rickard> EricDalquist: The db config you posted earlier (http://pastebin.com/p0g00z7G). Would you say that makes sense as a general default? As in better than the config that's default in 3.2?
[14:54:27 CDT(-0500)] <EricDalquist> probably
[14:54:58 CDT(-0500)] <EricDalquist> you might want to tweak 50# based on your # of servers, db size, etc
[14:55:13 CDT(-0500)] <EricDalquist> that config will grow the pool to the max needed conccurrent connections
[14:55:21 CDT(-0500)] <EricDalquist> and then slowly prune unused connections
[14:55:32 CDT(-0500)] <EricDalquist> so you'll see the pool size grow and shrink with usage
[14:55:56 CDT(-0500)] <athena> argh, today really just is full of weird problems
[14:56:12 CDT(-0500)] <athena> i'm now getting "javax.servlet.ServletException: File &amp;quot;/WEB-INF/jsp/ajax/jsonView.jsp&amp;quot; not found" for an ajax portlet on this linux server
[14:56:18 CDT(-0500)] <athena> no idea why - works fine locally
[14:56:25 CDT(-0500)] <athena> no idea why it think that's a JSP
[14:56:51 CDT(-0500)] <EricDalquist> view resolver ordering in the portlet
[14:57:50 CDT(-0500)] <rickard> EricDalquist: Did you guys remove removeAbandonded, removeAbandonedTimeout, logAbandoned and poolPreparedStatements or are those lines just missing from your paste?
[14:57:52 CDT(-0500)] <athena> yeah, but . . . it works elsewhere
[14:57:55 CDT(-0500)] * athena is going back to bed
[14:58:00 CDT(-0500)] <EricDalquist> we don't use them
[14:58:11 CDT(-0500)] <EricDalquist> but we haven't had problems with abandonded connections in a long while
[14:58:40 CDT(-0500)] <rickard> Ok. An abandoned connection is a connection that is opened, but never closed, right?
[14:58:59 CDT(-0500)] <EricDalquist> right
[14:59:04 CDT(-0500)] <rickard> Ok, got it. Thanks.
[14:59:12 CDT(-0500)] <EricDalquist> so its useful if you have a bad bit of code somwhere
[14:59:23 CDT(-0500)] <athena> feel like i've maybe seen those be more of a problem in environments where there isn't a great network connection between the portal and the db, too?
[15:00:14 CDT(-0500)] <rickard> "if you have a bad bit of code" - oh, I'm pretty sure we have that (smile)
[15:00:21 CDT(-0500)] <EricDalquist> (smile)
[15:00:46 CDT(-0500)] <rickard> From ealier employees, of course.
[15:27:24 CDT(-0500)] <athena> sometimes it's the driver's fault, too