uPortal IRC Logs-2011-01-28
- Apereo Infrastructure
[14:35:49 CST(-0600)] <dx9s_work> Is there known issues with MySQL !? I'm getting error messages from ant crn-import "ERROR Failed to import Structure from: structure/DLM_mobile_columns-2.structure due to exception: org.danann.cernunnos.ManagedException: The Cernunnos Runtime encountered an error:" as well as org.sprintframework.jdbc.BadSqlGrammerException PreparedStatmentCallback; bad SQL grammer ? <dx9s_work> shared.loader says to add shared.loader=$ /shared/classes,$ /shared/lib/*.jar and mine is shared.loader=$ /shared/classes,$ /shared/.jar,/var/lib/tomcat6/shared/classes,/var/lib/tomcat6/shared/.jar <parkerneff> JAVA_OPTS="$ -XX:MaxPermSize=192m" <holdorph> $ /uportal-war/src/main/resources/layout/theme/universality/components.xsl <dx9s_work> what/where is $ .. is that the source home or cataline.base ?
[14:36:07 CST(-0600)] <dx9s_work> ant dbtest passes..
[14:39:22 CST(-0600)] <dx9s_work> ant dbinit failed inside crn-import: http://uportal.pastebin.com/YHUUYrQb
[14:40:27 CST(-0600)] <holdorph> you've checked this page: https://wiki.jasig.org/display/UPM32/MySQL
[14:41:17 CST(-0600)] <dx9s_work> been there several times... and ant dbtest passes
[14:41:44 CST(-0600)] <holdorph> looks very very very similar to the error described in the red box on that page
[14:42:14 CST(-0600)] <holdorph> which is related to the "lower_case_table_names" option of mysql described in the yellow box above it
[14:43:24 CST(-0600)] <dx9s_work> I put that in the /etc/mysql/my.cnf ... I re-done it several times.. making sure it is there this time as well
[14:44:00 CST(-0600)] * dx9s_work quotes:
[14:44:01 CST(-0600)] <dx9s_work> [mysqld]
[14:44:01 CST(-0600)] <dx9s_work> #next few lines for tomcat6/uPortal
[14:44:01 CST(-0600)] <dx9s_work> default-storage-engine=innodb
[14:44:01 CST(-0600)] <dx9s_work> #lower_case_table_names=1
[14:44:01 CST(-0600)] <dx9s_work> #set-variable = innodb_buffer_pool_size=128M
[14:44:03 CST(-0600)] <dx9s_work> #set-variable = innodb_log_file_size=256M
[14:44:05 CST(-0600)] <dx9s_work> #set-variable = innodb_additional_mem_pool_size=20M
[14:44:07 CST(-0600)] <dx9s_work> innodb_flush_log_at_trx_commit=1
[14:44:09 CST(-0600)] * dx9s_work ends quote
[14:44:17 CST(-0600)] <holdorph> you can verify it's 'working' by trying to create a table through the mysql command line program
[14:44:17 CST(-0600)] <dx9s_work> I commented it out this time.. duh
[14:44:37 CST(-0600)] <holdorph> and then trying to insert into it with a different case syntax
[14:44:45 CST(-0600)] <dx9s_work> doing then un-doing.. then redoing.. bound to miss something
[14:44:51 CST(-0600)] <dx9s_work> (uportal should be this complicated)
[14:45:11 CST(-0600)] <holdorph> the case sensitivity thing is a mysql problem
[14:45:31 CST(-0600)] <holdorph> sql is not case sensitive, that should never have been a default behavior for a SQL compatible database
[14:48:33 CST(-0600)] <dx9s_work> one thing that I think is missing from here: https://wiki.jasig.org/display/UPM32/MySQL is to comment out the first database definition in /uPortal/uportal-impl/src/main/resources/properties/rdbm.properties or not.. it ap that you should have: pears
[14:48:35 CST(-0600)] * dx9s_work quotes:
[14:48:36 CST(-0600)] <dx9s_work> ##### Hypersonic SQL - Server mode
[14:48:36 CST(-0600)] <dx9s_work> ##### Requires first starting Hypersonic SQL with the command:
[14:48:36 CST(-0600)] <dx9s_work> ##### "ant hsql"
[14:48:36 CST(-0600)] <dx9s_work> #hibernate.connection.driver_class=org.hsqldb.jdbcDriver
[14:48:37 CST(-0600)] <dx9s_work> #hibernate.connection.url=jdbc:hsqldb:hsql://localhost:8887
[14:48:39 CST(-0600)] <dx9s_work> #hibernate.connection.username=sa
[14:48:41 CST(-0600)] <dx9s_work> #hibernate.connection.password=
[14:48:43 CST(-0600)] <dx9s_work> #hibernate.dialect=org.hibernate.dialect.HSQLDialect
[14:48:45 CST(-0600)] * dx9s_work end quotes
[14:48:47 CST(-0600)] <dx9s_work> commented out
[14:49:30 CST(-0600)] <dx9s_work> trying AGAIN
[14:49:54 CST(-0600)] <holdorph> it would probably be good to be 100% clear on that. but if you're familiar with java applications and .properties files it's a pretty well known format
[14:50:08 CST(-0600)] <holdorph> two properties of the same name in a .properties file, the last one wins
[14:50:43 CST(-0600)] * dx9s_work is crossing fingers
[14:50:57 CST(-0600)] <dx9s_work> (inside initdb now)
[14:52:06 CST(-0600)] <dx9s_work> DUH!... all because I undid what I previously did and forgot to re-do it!
[14:52:48 CST(-0600)] <dx9s_work> welp I got it working with system packages – have to be a little creative in /etc/environment and created a few sym links, but got it working with system packages inside a debian distro (ubuntu)
[14:53:02 CST(-0600)] <dx9s_work> well the initportal part.. now going to attempt deploying
[14:53:31 CST(-0600)] <dx9s_work> holdorph, thanks for putting up w/ me
[14:53:40 CST(-0600)] <holdorph> np
[14:54:28 CST(-0600)] <dx9s_work> if I get around to it.. I might write up a howto on ubuntu with 3.2.4 and using system ant/tomcat/etc... in "System" locations
[14:54:54 CST(-0600)] <dx9s_work> so what is safer... deploy-war or deploy-ear ?!
[14:55:05 CST(-0600)] <dx9s_work> I figure ear is more complete
[14:55:13 CST(-0600)] <holdorph> you can't do only deploy-war the first time
[14:55:36 CST(-0600)] <holdorph> deploy-war will only deploy the uportal web application, not any of the other supporting web apps (like portlets)
[14:56:13 CST(-0600)] <holdorph> so, unless you've completely cleaned out your uportal to not have any portlets (like webproxy for instance) you'll have to do deploy-ear at least once
[14:56:20 CST(-0600)] <dx9s_work> oh the initportal deployed it already?
[14:56:52 CST(-0600)] <holdorph> initportal is a parent target that calls a LOT of secondary targets. and yeah, I think deploy-ear is one of those.
[14:57:01 CST(-0600)] <holdorph> it's in the build.xml if you take a look.
[14:58:44 CST(-0600)] <dx9s_work> hmm.. see stuff in webapps.. but restarting tomcat... need to look at log files
[15:03:22 CST(-0600)] <dx9s_work> something up with creating a springframework brean ...
[15:03:57 CST(-0600)] <holdorph> all of the shared class libraries are in the right place and the config file for it is set up correctly?
[15:04:31 CST(-0600)] <dx9s_work> that was a question I had... I might have to poke around with shared.loaded ..
[15:04:35 CST(-0600)] <holdorph> deploy-ear should put the shared files in place to a normal tomcat, but if you're using a system install toomcat, they might not have gone to the right place
[15:04:58 CST(-0600)] <holdorph> https://wiki.jasig.org/display/UPM32/Installing+Tomcat
[15:05:06 CST(-0600)] <holdorph> specifically the 'Shared libraries' section
[15:06:06 CST(-0600)] <holdorph> shared/lib$ ls
[15:06:06 CST(-0600)] <holdorph> castor-1.1.1.jar pluto-descriptor-api-1.1.7.jar
[15:06:06 CST(-0600)] <holdorph> commons-logging-api-1.1.jar pluto-descriptor-impl-1.1.7.jar
[15:06:06 CST(-0600)] <holdorph> person-directory-api-1.5.0-RC6.jar pluto-taglib-1.1.7.jar
[15:06:06 CST(-0600)] <holdorph> pluto-container-1.1.7.jar portlet-api-1.0.jar
[15:06:11 CST(-0600)]
[15:06:24 CST(-0600)] <dx9s_work> base is /var/lib/tomcat6
[15:06:57 CST(-0600)] <holdorph> those jars i pasted above need to be in one of the directories mentioned in the shared.loader line
[15:06:59 CST(-0600)] <dx9s_work> I need to poke around and see where that framework bean is and verify that it is in the path
[15:13:44 CST(-0600)] <dx9s_work> too many locations
[15:20:02 CST(-0600)] <dx9s_work> I found that maven2 system packet makes a /usr/shared/maven2 which is the say layout as if you has installed in into home... so a system M2_HOME="/usr/shared/maven2" M2="/usr/shared/maven2/bin" (those of which the package maintains symlinks to /usr/bin/mvn and so forth... as for tomcat.. put in a few links in /var/lib/tomcat to point to /usr/shared/tomcat/... (which itself has conf point to /etc/tomcat6 and a few other things)... symlinks to symlinks
[15:20:02 CST(-0600)] <dx9s_work> .. but /var/lib/tomcat is as-if one had installed it into their home
[15:20:37 CST(-0600)] <dx9s_work> then simply TOMCAT_HOME=/var/lib/tomcat6 ...
[15:21:43 CST(-0600)] <dx9s_work> but localhost:8080/uPortal/ is just sitting there... (updated shared.loaded to include /var/lib/tomcat6/shared/lib/*.jar ) ... the beans are found... next error message
[15:21:48 CST(-0600)] <dx9s_work> (need to find it)
[15:22:33 CST(-0600)] <dx9s_work> hmm.. something about car
[15:24:32 CST(-0600)] <dx9s_work> any clues... "INFO [uPortalTaskScheduler_Worker-4] Jan/28 13:16:13,867 car.CarResources.[] - Channel Archives can not be loaded. CAR directory '/var/lib/tomcat6/webapps/uPortal/WEB-INF/cars' does not exist. "
[15:31:37 CST(-0600)] <holdorph> that is not a problem, that's why it's only an INFO line
[15:31:59 CST(-0600)] <holdorph> it shows up in even running uportal installs, as long as there aren't any 'car' files
[15:32:59 CST(-0600)] <dx9s_work> oh.. it appears to be a memory tweak I need to do w/ tomcat
[15:33:10 CST(-0600)] <dx9s_work> (increase some memory) – I thought I saw that someplace
[15:33:47 CST(-0600)] <dx9s_work> https://wiki.jasig.org/display/UPC/uPortal+Heap+Tuning
[15:34:04 CST(-0600)] <holdorph> you definitely have to increase permgen with uportal 3.x
[15:34:21 CST(-0600)] <holdorph> the default permgen size of jdk 1.5 and 1.6 won't cut it
[15:35:57 CST(-0600)] <dx9s_work> what is the more "correct" way to increase heap and 'permgen' ? from a conf/* file ?
[15:36:07 CST(-0600)] <dx9s_work> or the init.d/tomcat6 startup script?
[15:36:30 CST(-0600)] <holdorph> you should have someplace where you pass java options to tomcat
[15:36:41 CST(-0600)] <holdorph> you'll need a whole lot more then just permgen for a production system anyway
[15:36:51 CST(-0600)] <holdorph> you can get away with only permgen setting for a developer box
[15:37:04 CST(-0600)] <dx9s_work> what do you pass ?
[15:37:51 CST(-0600)] <holdorph> typically start memory, max memory, permgen, permgen gc, the normal memory gc algorithm, ..... and sometimes more
[15:38:06 CST(-0600)] <holdorph> long time system admins tend to build them up over time.
[15:38:47 CST(-0600)] <dx9s_work> any specific values?
[15:39:11 CST(-0600)] <parkerneff> Assuming Ubuntu Server edit /etc/default/tomcat6 and add
[15:39:18 CST(-0600)]
[15:39:33 CST(-0600)] <holdorph> the values are dependent on the system
[15:39:39 CST(-0600)] <holdorph> there are LOTS AND LOTS of factors
[15:39:54 CST(-0600)] <holdorph> how many portlet,, how many layouts, how many users, and on and on
[15:41:13 CST(-0600)] <dx9s_work> so no golden rule of 192 or 256 is a "safe" number to use !?
[15:42:25 CST(-0600)] <holdorph> a lot of it depends on how many different portlet web applications you have
[15:42:34 CST(-0600)] <holdorph> the more portlet web apps, the more permgen you'll need
[15:42:49 CST(-0600)] <holdorph> i think 192 is a good minimum for the default of what ships with uportal out of the box
[15:43:03 CST(-0600)] <holdorph> but if you add to that, without taking any away, then you might want to increase it
[15:43:45 CST(-0600)] <dx9s_work> I set to 256 for now
[15:44:29 CST(-0600)] <dx9s_work> seems to be happy
[15:46:11 CST(-0600)] <dx9s_work> hmmm.. is it normal for (once logged in as admin) that the bookmarks things to have an "error report" and some issue with PortletDD ?
[15:47:38 CST(-0600)] <dx9s_work> the trace is has this for a line: org.jasig.portal.channels.portlet.InconsistentPortletModelException: Could not retrieve PortletDD for portlet window '40_u24l1n8_8.u24l1n8', this usually means the Portlet application is not deployed correctly.
[15:47:45 CST(-0600)] <holdorph> nope.
[15:48:27 CST(-0600)] <holdorph> if you're dealing with one of the shipps-with-uportal portlets, the problem is probably the portlet web application for that portlet, not deploying (or starting) correctly
[15:48:43 CST(-0600)] <holdorph> check to make sure that portlet web app is present in the tomcat webapps directory
[15:51:34 CST(-0600)] <dx9s_work> hmmm. nothing: setup@uportal:/var/lib/tomcat6/webapps$ find /var/lib/tomcat6 | egrep -i '(portletdd|40_u24)'
[15:51:34 CST(-0600)] <dx9s_work> setup@uportal:/var/lib/tomcat6/webapps$ find /home/setup/uPortal-3.2.4 | egrep -i '(portletdd|40_u24)'
[15:52:01 CST(-0600)] <holdorph> why are you grepping on that?
[15:52:04 CST(-0600)] <dx9s_work> it must just be something that wasn't included in the uPortal source
[15:52:07 CST(-0600)] <holdorph> that's just an internal uportal id
[15:52:40 CST(-0600)] <holdorph> $ ls webapps/
[15:52:40 CST(-0600)] <holdorph> BookmarksPortlet host-manager RssPortlet
[15:52:40 CST(-0600)] <holdorph> cas jasig-widget-portlets uPortal
[15:52:40 CST(-0600)] <holdorph> cas-proxy-test-portlet manager WeatherPortlet
[15:52:40 CST(-0600)] <holdorph> docs pluto-testsuite WebProxyPortlet
[15:52:40 CST(-0600)] <holdorph> examples ResourceServingWebapp WebUtils
[15:52:42 CST(-0600)] <holdorph> FunctionalTestsPortlet ROOT
[15:53:06 CST(-0600)] <dx9s_work> a way to check all files from a point.. not just top level
[15:53:39 CST(-0600)] <holdorph> i just mean the value of '(portletdd|40_u24)' is never going to hit anythign on the filesystem
[15:53:54 CST(-0600)] <holdorph> it's an internal uportal generated id. it's not an id in the portlet application
[15:53:59 CST(-0600)] <dx9s_work> another way grep -irl portletdd *
[15:54:10 CST(-0600)] <dx9s_work> that shows stuff
[15:54:32 CST(-0600)] * dx9s_work is new to this whole "portlet" thing...
[15:54:49 CST(-0600)] <holdorph> what makes up a portlet deployment descriptor externally, at the tomcat level is
[15:54:53 CST(-0600)] <holdorph> the portlet web application name
[15:55:02 CST(-0600)] <holdorph> and the portlet-name from the portlet.xml file
[15:55:26 CST(-0600)] <holdorph> and that is only 'realized' in the tomcat container, if there is the 'pluto goo' in the web.xml for that portlet-name
[15:56:28 CST(-0600)] <dx9s_work> there are lots of portlet.xml files in base
[15:59:57 CST(-0600)] <holdorph> ?
[16:01:05 CST(-0600)] <dx9s_work> just trying to understand how this whole thing works...
[16:02:33 CST(-0600)] <holdorph> i just don't know what you meant by "files in base"
[16:02:43 CST(-0600)] <dx9s_work> $(catalina.base}
[16:02:50 CST(-0600)] <dx9s_work> aka /var/lib/tomcat6
[16:03:01 CST(-0600)] <holdorph> yes, there should be one portlet.xml file per portlet web application
[16:03:21 CST(-0600)] <holdorph> and in a uportal tomcat, nearly every web application will be a portlet web application
[16:03:40 CST(-0600)] <dx9s_work> what is 'PortletDD' ?
[16:04:31 CST(-0600)] <holdorph> that stands for portlet deployment descriptor. which is a uportal identifier
[16:05:00 CST(-0600)] <holdorph> that is (generated and) associated with the information of the portlet web application, web app id, and the portlet-name
[16:05:52 CST(-0600)] <dx9s_work> so "org.jasig.portal.channels.portlet.InconsistentPortletModelException: Could not retrieve PortletDD for portlet window '40_u24l1n8_8.u24l1n8', this usually means the Portlet application is not deployed correctly." .. I have to examine the window with that ID (40_u24l1n8_8.u24l1n8)
[16:05:55 CST(-0600)] <dx9s_work> ?
[16:06:18 CST(-0600)] <holdorph> ignore the id for a moment
[16:06:25 CST(-0600)] <holdorph> you're having a problem with bookmarks
[16:06:38 CST(-0600)] <holdorph> do you see the bookmarks web application in the tomcat webapps directory?
[16:07:13 CST(-0600)] <dx9s_work> yes
[16:07:38 CST(-0600)] <dx9s_work> assuming BookmarksPortlet is it
[16:08:14 CST(-0600)] <dx9s_work> "setup@uportal:/var/lib/tomcat6/webapps$ ls
[16:08:14 CST(-0600)] <dx9s_work> BookmarksPortlet cas cas-proxy-test-portlet FunctionalTestsPortlet jasig-widget-portlets pluto-testsuite ResourceServingWebapp ROOT RssPortlet uPortal WeatherPortlet WebProxyPortlet"
[16:08:17 CST(-0600)] <holdorph> were there any problems with that web app starting that you noticed during start up?
[16:08:37 CST(-0600)] <holdorph> i am guessing because of your non-standard (from uportal point of view) tomcat directory layout
[16:08:48 CST(-0600)] <holdorph> that it's having a hard time finding ANY of the portlet web applications
[16:08:49 CST(-0600)] <dx9s_work> yeah.. it /var/log/tomcat6/BookmarksPortlet.log ...
[16:08:55 CST(-0600)] <holdorph> it has to read the portlet.xml file
[16:09:00 CST(-0600)] * dx9s_work is reading that log now
[16:09:12 CST(-0600)] <holdorph> and if it can't find those web apps, it can't read that file
[16:10:53 CST(-0600)] <dx9s_work> the portlet.xml is there.. the log has this one liner: ERROR [main] Jan/28 13:42:58,785 context.ContextLoader.[] - Context initialization failed (and stuff after it... still reading)
[16:11:17 CST(-0600)] <dx9s_work> arg.. another bean issue with tranactionManager
[16:11:22 CST(-0600)] <holdorph> sounds like the app didn't start then.
[16:13:04 CST(-0600)] <dx9s_work> working down the trace... it might be JDBC related
[16:13:52 CST(-0600)] * dx9s_work pastes: http://uportal.pastebin.com/Vr9TtaEz
[16:15:12 CST(-0600)] <dx9s_work> I am guessing I have to setup a JDBC connection for it as well
[16:15:53 CST(-0600)] <holdorph> yes, but that should have been handled for you if you had configured the database correctly before doing the 'initportal'
[16:17:07 CST(-0600)] <dx9s_work> magic word: should... I did clear out the DB and start over with a new initportal and it deployed and all that.. perhaps I need to remove from my .m2 cache the bookmarkportlet as it might have cached bad DB stuff
[16:17:50 CST(-0600)] <holdorph> did you do a clean, before the init?
[16:18:25 CST(-0600)] <dx9s_work> what part .. I did a drop/create on the database before the ant initportal (which it should also do according to docs)... what else to clean out?
[16:19:01 CST(-0600)] <dx9s_work> ant clean ?
[16:19:12 CST(-0600)] <holdorph> yes, ant clean
[16:19:57 CST(-0600)] <parkerneff> I had a similar issue with the Bookmark portlet.. We were using postgresql and the Bookmark portlet couldnt find the driver. I had to put the postgres jdbc driver jar in the WEB-INF/lib of the bookmark portlet
[16:20:50 CST(-0600)] <holdorph> hmm.... haven't noticed, that before, but I seldom do installs with anything besides HSQL
[16:21:01 CST(-0600)] <holdorph> and for those that I do, i might not have checked out bookmarks
[16:21:36 CST(-0600)] <dx9s_work> did-done: service tomcat6 stop; ant clean; mysql (drop/create); ... doing: ant initportal
[16:21:42 CST(-0600)] <holdorph> yeah, i see hsqldb.*.jar already in the bookmarks web-inf/lib directory
[16:23:32 CST(-0600)] * dx9s_work adds (crossing fingers)..
[16:24:07 CST(-0600)] <holdorph> if you didn't take parkerneff 's suggestion, i guess based on his experience and what I can see in the directory, you'll probably still have a problem
[16:24:52 CST(-0600)] <dx9s_work> you might be a good guesser...
[16:25:33 CST(-0600)] * dx9s_work is imagining the "Jeopardy" final round theme music ...
[16:26:51 CST(-0600)] <holdorph> it might be worth doing an ls on the BookmarksPortlet/WEB-INF/lib directory before starting up and seeing if there is a mysql jar file in there
[16:26:52 CST(-0600)] <dx9s_work> holdorph, what did you wadger!?
[16:27:19 CST(-0600)] <holdorph> 
no one called my bet.
[16:27:40 CST(-0600)] <dx9s_work> if you had wadgered your full amount... you would have doubled your score!
[16:28:40 CST(-0600)] <dx9s_work> setup@uportal:/var/lib/tomcat6/webapps/BookmarksPortlet/WEB-INF/lib$ ls
[16:28:40 CST(-0600)] <dx9s_work> activation-1.1.jar cernunnos-1.1.0.jar commons-jexl-1.1.jar groovy-engine-1.1-jdk14.jar js-engine-1.1-jdk14.jar spring-aop-2.0.5.jar spring-ldap-1.2.1.jar xercesImpl-2.6.2.jar
[16:28:40 CST(-0600)] <dx9s_work> ant-1.7.0.jar cglib-2.1_3.jar commons-lang-2.3.jar hibernate-3.2.3.ga.jar jstl-1.1.2.jar spring-beans-2.0.5.jar spring-portlet-2.0.5.jar xml-apis-1.0.b2.jar
[16:28:40 CST(-0600)] <dx9s_work> ant-launcher-1.7.0.jar commons-codec-1.2.jar commons-logging-1.1.jar hsqldb-1.8.0.7.jar jta-1.0.1B.jar spring-context-2.0.5.jar spring-support-2.0.5.jar xmlParserAPIs-2.6.2.jar
[16:28:45 CST(-0600)] <dx9s_work> antlr-2.7.6.jar commons-collections-3.2.jar commons-pool-1.3.jar icu4j-2.6.1.jar log4j-1.2.14.jar spring-core-2.0.5.jar spring-tx-2.5.5.jar xom-1.0.jar
[16:28:48 CST(-0600)] <dx9s_work> aopalliance-1.0.jar commons-dbcp-1.2.2.jar cvsclient-20060125.jar jaxen-1.1.1.jar mail-1.4.1.jar spring-dao-2.0.5.jar spring-web-2.0.5.jar
[16:28:51 CST(-0600)] <dx9s_work> asm-1.5.3.jar commons-fileupload-1.2.1.jar dom4j-1.6.1.jar jdom-1.0.jar nekohtml-1.9.6.2.jar spring-hibernate3-2.0.5.jar spring-webmvc-2.0.5.jar
[16:28:55 CST(-0600)] <dx9s_work> asm-attrs-1.5.3.jar commons-httpclient-3.0.1.jar ehcache-1.2.3.jar jline-0.9.94.jar retrotranslator-runtime-1.2.7.jar spring-jdbc-2.0.5.jar standard-1.1.2.jar
[16:28:58 CST(-0600)] <dx9s_work> backport-util-concurrent-3.1.jar commons-io-1.4.jar groovy-all-1.5.6.jar js-1.7R1.jar script-api-1.0.jar spring-jmx-2.0.5.jar xalan-2.6.0.jar
[16:29:59 CST(-0600)] <dx9s_work> so what... just copy (or link) the mysql-connector-java-5.1.6.jar ?
[16:30:21 CST(-0600)] <holdorph> you can copy for now and try to restart.
[16:30:40 CST(-0600)] <holdorph> but to fix the problem for new initportal/deploy-ear runs, you'd have to do something different
[16:31:24 CST(-0600)] <dx9s_work> that was it... hmmm
[16:31:45 CST(-0600)] <holdorph> i'm surprised it didn't pick it up though. looking at the uportal source tree, in this file:
[16:32:04 CST(-0600)] <holdorph> uPortal-3.2.x/uportal-portlets-overlay/BookmarksPortlet/pom.xml
[16:32:14 CST(-0600)] <holdorph> you'll see a maven dependency for the jdbc driver.
[16:32:39 CST(-0600)] <holdorph> it's based on maven properties that should have been set by the parent (uportal) project
[16:32:41 CST(-0600)] <athena> any chance the configured jdbc driver in the main pom.xml was edited after the initial initportal?
[16:32:57 CST(-0600)] <dx9s_work> it's a bug that has to do with the phase of the moon in the constellation Orion! I mean really..ANYBODY would know THAT (j/k)
[16:32:59 CST(-0600)] <athena> if so, you'd need to run a clean, then a full deploy-ear to get it to be updated
[16:36:24 CST(-0600)] <dx9s_work> I haven't touched the main pom file since the initial edit per the install/build instructions: https://wiki.jasig.org/pages/viewpage.action?pageId=29133275 – really the POM edits listed from https://wiki.jasig.org/display/UPM32/MySQL
[16:38:01 CST(-0600)] <dx9s_work> I did ant clean; ant initportal (of which deploy-ear is part of it) ; however does it NOT update things already deployed? aka I have to remove from webapps the bookmark portlet for a deploy-ear to install a correct one?
[16:38:42 CST(-0600)] <dx9s_work> or does deploy-ear overwrite what is there?
[16:40:14 CST(-0600)] <dx9s_work> next steps.. look into external authentication (ldap or kerb or radius, which ever one is easier).. assuming uPortal support several authentication sources (local DB, external LDAP, etc.) so the admin account can be a local one !?
[16:40:41 CST(-0600)] <holdorph> the default on initportal, is to remove existing directories
[16:41:18 CST(-0600)] <holdorph> the out-of-the-box uportal config is configured to authenticate against CAS
[16:41:41 CST(-0600)] <holdorph> so you have to decide if you want to configure CAS to authenticate against those other sources
[16:41:42 CST(-0600)] <dx9s_work> and I've messed with CAS last year.. got it to work off of LDAP I think is what I did.
[16:41:57 CST(-0600)] <holdorph> or remove CAS and have uportal authenticate directly against the sources like ldap
[16:42:18 CST(-0600)] <dx9s_work> holdorph, in case I forget to say it.. thanks..
[16:42:30 CST(-0600)] <holdorph> yw
[16:44:16 CST(-0600)] <dx9s_work> the question is which is better.. if I can point CAS at RADIUS, I can do the hand-off to the various (kerb) realms there instead of in CAS... I remember some rewrite rule on looking up and hacking so it looked like the person was doing an email address... (yuk in any case, the rewriting, not the "email" address part)
[16:44:30 CST(-0600)] <dx9s_work> I forget what I did tho...
[16:45:10 CST(-0600)] <dx9s_work> but I wanted to go to uPortal and saw that CAS was involved.. so I worked on CAS directly.. then I never got around to uPortal.. hehe now I forgot the CAS hacks I did.
[16:45:14 CST(-0600)] <holdorph> simplest, I think is pointing cas at ldap and then pointing uportal at ldap for person attributes
[16:46:30 CST(-0600)] <dx9s_work> I need to fold several "realms" (be it ldap, kerb, or radius as a middle man to kerb) into one... no centralized user accounting here... (no ONE user ldap DB)
[16:49:17 CST(-0600)] <holdorph> well you have options, you can let uportal authenticate against cas and do the work in cas. or you can remove the cas and let uportal authenticate directly off the ldap server. I'm not sure if uportal ldap authentication supports multiple servers or not. using cas is probably more common then that, i expect
[16:50:08 CST(-0600)] <dx9s_work> does uportal (if directly doing authentication) support multiple sources !? as for cas and two-or-more LDAP sources. it only looks like one.
[16:50:24 CST(-0600)] <dx9s_work> I am guessing yes
[16:51:16 CST(-0600)] <holdorph> yes, uportal supports looking up multiple sources, i've had it using CAS, direct-ldap, and db all at the same time
[16:51:27 CST(-0600)] <dx9s_work> sweet!
[16:54:13 CST(-0600)] <dx9s_work> I see this.. https://wiki.jasig.org/display/UPM32/Authentication is there a better/alternative source that specifically covers multi-sourced authentication
[16:56:26 CST(-0600)] <holdorph> it has some good information in it. it's a combination of things though. one part is the configuring of uportal-impl/src/main/resources/properties/security.properties
[16:56:32 CST(-0600)] <holdorph> like you see in https://wiki.jasig.org/display/UPM32/Authenticating+Against+LDAP
[16:57:04 CST(-0600)] <holdorph> and then if you need to get username/pw field into uportal for one of those options, you'll have to adjust the theme
[16:57:07 CST(-0600)] * holdorph looks for that
[16:58:17 CST(-0600)] <holdorph> athena, probably remembers..... (i have to hunt for it)
[16:58:40 CST(-0600)] * athena looks up
[16:58:51 CST(-0600)] <athena> yes, uportal supports multiple authentication sources
[16:58:59 CST(-0600)] <holdorph> just how do you turn on the username/pw field, if you want to authenticate against another source
[16:59:19 CST(-0600)] <athena> ah.
[16:59:39 CST(-0600)] <athena> i think the current functionality disables the username/pw field when cas is enabled
[17:00:08 CST(-0600)] <athena> so i think if you uncomment all the CAS-related lines in security.properties the username/password fields will be displayed again
[17:00:52 CST(-0600)] <athena> though i think we'd generally recommend against disabling CAS
[17:01:01 CST(-0600)] <athena> are you looking to use uportal's internal account store or ?
[17:01:02 CST(-0600)]
[17:01:12 CST(-0600)] <holdorph> search for TEMPLATE: LOGIN
[17:01:42 CST(-0600)] <holdorph> you can just adjust (remove) the decision making code if you want both active
[17:01:44 CST(-0600)] <athena> scrolling up a bit, i think i'd really recommend using CAS rather than doing LDAP directly through uPortal
[17:01:52 CST(-0600)]
[17:01:52 CST(-0600)] <holdorph> (the logon with cas button and the username/pw fields)
[17:02:06 CST(-0600)] <holdorph> the root of the source tree
[17:04:57 CST(-0600)] <dx9s_work> hmmm the external.login template creates the html fields with CAS in the name... so is THAT why you don't recommend disabling CAS ?
[17:05:59 CST(-0600)] <holdorph> no, just that CAS has spent a lot more time fine tuning the 'authenticate against ldap' code then uportal has
[17:06:34 CST(-0600)] <holdorph> at least that's what I assume athena meant
[17:07:08 CST(-0600)] <athena> yes, that's correct
[17:07:16 CST(-0600)] <dx9s_work> yeah.. If I can get Radius to work (which shouldn't be that hard)... I think letting it steer the authentication to the correct "realm" might work fair enough... -> https://wiki.jasig.org/display/CASUM/RADIUS
[17:07:39 CST(-0600)] <athena> and while i think you could authenticate against multiple LDAP stores from uportal, you're not going to find documentation for it, and it'll be much easier through CAS
[17:07:42 CST(-0600)] <dx9s_work> (I will retract that... it could be hard)
[17:08:10 CST(-0600)] <athena> and in the end, uportal generally defers to CAS on authentication tasks - i think uPortal's LDAP authentication code could be considered close to deprecated at this point
[17:08:36 CST(-0600)] <dx9s_work> so CAS is best anyways (LDAP/Kerb/etc.)
[17:09:28 CST(-0600)] <athena> yes
[17:09:35 CST(-0600)] <athena> CAS is good.
[17:10:35 CST(-0600)] <dx9s_work> I will probably ask (stupid even) question about hacking CAS changes into the uPortal source.. but not today, you folks get a pseudo-beer today!
[17:11:53 CST(-0600)] <holdorph> woohoo! it's beer-30 somewhere in the world i'm sure.
[17:12:04 CST(-0600)] <dx9s_work> heh
[17:13:31 CST(-0600)] <dx9s_work> from what I understand webapps/cas is CAS... webapps/cas-proxy-text-portlet is a uportal portlet to access cas ... things are beginning to fall into place... almost like (A-Team, Hannibal Smith) 'I love it when a plan comes together"
[17:14:16 CST(-0600)] <athena> excellent, glad it feels like it
[17:14:33 CST(-0600)] <athena> to clarify, the cas-proxy-test-portlet is just a test resource that demonstrates how portlets might request proxy tickets from cas
[17:14:42 CST(-0600)] <athena> but isn't at all required to do most cas authentication