Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 36 Next »

[04:40:10 EST(-0500)] * colinclark (n=colin@adsl-69-110-30-151.dsl.pltn13.pacbell.net) has joined ##uportal
[07:29:57 EST(-0500)] * jayshao (n=jayshao@pool-72-79-114-2.nwrknj.east.verizon.net) has joined ##uportal
[09:04:37 EST(-0500)] * esm (n=esm@128.220.212.115) has joined ##uportal
[09:49:33 EST(-0500)] * anastasiac (n=team@142.150.154.149) has joined ##uportal
[10:01:23 EST(-0500)] * EricDalquist (n=dalquist@bohemia.doit.wisc.edu) has joined ##uportal
[10:01:27 EST(-0500)] * theclown (n=clown@142.150.154.101) has joined ##uportal
[10:09:45 EST(-0500)] * dstn (n=dstn@unaffiliated/dstn) has joined ##uportal
[10:25:10 EST(-0500)] * athena7 (n=athena7@lumina.its.yale.edu) has joined ##uportal
[11:02:05 EST(-0500)] * colinclark (n=colin@adsl-69-110-30-151.dsl.pltn13.pacbell.net) has joined ##uportal
[11:25:41 EST(-0500)] * michelled (n=team@142.150.154.199) has joined ##uportal
[12:52:53 EST(-0500)] * colinclark (n=colin@srb-wlan-56.AirBears.Berkeley.EDU) has joined ##uportal
[13:56:01 EST(-0500)] * esm (n=esm@128.220.212.115) has joined ##uportal
[14:15:43 EST(-0500)] * colinclark (n=colin@srb-wlan-56.AirBears.Berkeley.EDU) has joined ##uportal
[14:19:07 EST(-0500)] * athena7 (n=athena7@lumina.its.yale.edu) has joined ##uportal
[14:53:10 EST(-0500)] * agherna (n=argherna@cites-agherna01.ci.uiuc.edu) has joined ##uportal
[15:05:36 EST(-0500)] * esm (n=esm@128.220.212.115) has joined ##uportal
[15:36:04 EST(-0500)] <dstn> Occasionally we receive an ArrayIndexOutOfBounds error from the portal causing random portlets to fail rendering. I was wondering if anyone else has experienced this problem. Here is the stacktrace: http://uportal.pastebin.org/22672
[15:37:11 EST(-0500)] <EricDalquist> interesting ... looks like the channel title incorporation is trying to get a title for a channel that channel manager doesn't know about ...
[15:37:18 EST(-0500)] <EricDalquist> no idea why that would be happening though (smile)
[15:39:37 EST(-0500)] * colinclark (n=colin@srb-wlan-34.AirBears.Berkeley.EDU) has joined ##uportal
[15:40:26 EST(-0500)] <dstn> Its weird, its so random when it happens too.
[15:44:54 EST(-0500)] <agherna> EricDalquist: so I have been working on this maximum open cursors issue that's been on the uportal-dev list
[15:45:02 EST(-0500)] <EricDalquist> yup
[15:45:05 EST(-0500)] <EricDalquist> saw that
[15:45:13 EST(-0500)] <agherna> i was able to look at our default settings
[15:45:21 EST(-0500)] <agherna> via jconsole/JMX
[15:45:23 EST(-0500)] <EricDalquist> we run on Oracle as well (10.something) and don't have any problems
[15:45:30 EST(-0500)] <agherna> that's good
[15:45:43 EST(-0500)] <agherna> are you guys using DLM?
[15:45:45 EST(-0500)] <EricDalquist> yup
[15:45:48 EST(-0500)] <agherna> ok
[15:46:05 EST(-0500)] <agherna> did you change the settings in the uPortal55.xml file at all?
[15:46:21 EST(-0500)] <EricDalquist> yeah
[15:46:21 EST(-0500)] <agherna> the defaults, I mean?
[15:46:30 EST(-0500)] <EricDalquist> we have fairly tweaked configs
[15:46:55 EST(-0500)] <agherna> I see.
[15:47:02 EST(-0500)] <agherna> what did you change?
[15:47:18 EST(-0500)] <EricDalquist> http://uportal.pastebin.com/d1a0f8f54
[15:47:30 EST(-0500)] * agherna scans pastbin
[15:47:59 EST(-0500)] <EricDalquist> I'm not sure any of the dbcp settings would help with a Statement/ResultSet leak
[15:48:12 EST(-0500)] <EricDalquist> do you have any customizations at all to any code that connects to the uPortal DB?
[15:48:28 EST(-0500)] <agherna> Hmmm... I don't think so.
[15:48:34 EST(-0500)] <agherna> what does this bit do: oracle.net.CONNECT_TIMEOUT=2000;oracle.net.READ_TIMEOUT=30000
[15:49:15 EST(-0500)] <EricDalquist> so that sets a 2 second timeout when the ojdbc driver calls Socket.connect
[15:49:21 EST(-0500)] <EricDalquist> and a 30 second timeout for Socket.read
[15:49:44 EST(-0500)] <EricDalquist> which fixes problems with the jdbc driver hanging if the database server isn't listening or failing to respond to read calls
[15:50:02 EST(-0500)] <EricDalquist> on the PortalDB datasource it doesn't matter much because either way the portal would be down
[15:50:19 EST(-0500)] <EricDalquist> for our other data sources it does matter because if they hang they can take the portal down
[15:50:33 EST(-0500)] <EricDalquist> even though the framework doesn't depend on those datasources, only portlets do
[15:51:15 EST(-0500)] <agherna> ij
[15:51:18 EST(-0500)] <agherna> er, ok.
[15:51:22 EST(-0500)] <agherna> home row...
[15:51:49 EST(-0500)] <dstn> lol
[15:52:23 EST(-0500)] <agherna> it really does matter where you put your hands...
[15:53:08 EST(-0500)] <agherna> hmm....
[15:54:13 EST(-0500)] <agherna> which oracle jdbc jar are you guys using?
[15:54:17 EST(-0500)] <agherna> classes12.jar?
[15:54:35 EST(-0500)] <dstn> That's what we use.
[15:54:42 EST(-0500)] <EricDalquist> 10.2.0.3.0
[15:54:51 EST(-0500)] <EricDalquist> ojdbc14.jar
[15:55:08 EST(-0500)] <agherna> does it make a difference?
[15:55:30 EST(-0500)] <EricDalquist> well it should match your database I assume
[15:55:40 EST(-0500)] <EricDalquist> so what version of oracle do you run?
[15:56:16 EST(-0500)] <agherna> Oracle Database 10g Enterprise Edition Release 10.2.0.2.0 - 64bit Production
[15:56:16 EST(-0500)] <agherna> With the Partitioning, OLAP and Data Mining options
[15:56:19 EST(-0500)] * michelled (n=team@142.150.154.199) has left ##uportal
[15:57:28 EST(-0500)] <agherna> i'm not a database guy, but it looks like it's reasonable new
[15:57:46 EST(-0500)] <EricDalquist> yeah
[15:57:54 EST(-0500)] <EricDalquist> you should probably be on a newer jdbc driver
[15:58:03 EST(-0500)] <EricDalquist> classes12.jar is the JDK1.2 compatible driver
[15:58:10 EST(-0500)] <agherna> yikes
[15:58:11 EST(-0500)] <agherna> that's old
[15:58:16 EST(-0500)] <agherna> really old
[15:58:26 EST(-0500)] <EricDalquist> do you know which version of classes12 you're using?
[15:58:42 EST(-0500)] <EricDalquist> oracle doesn't include the actual driver version in the jar name
[15:58:46 EST(-0500)] <agherna> checking
[15:59:45 EST(-0500)] <agherna> from the manifest: Implementation-Version: "Oracle JDBC Driver version - 10.1.0.2.0"
[16:00:17 EST(-0500)] <agherna> but
[16:00:30 EST(-0500)] <EricDalquist> so that is fairly recent
[16:00:31 EST(-0500)] <agherna> there's also Specification-Title: "Oracle JDBC driver classes for use with JDK1.
[16:00:31 EST(-0500)] <agherna> 2 and JDK1.3"
[16:00:35 EST(-0500)] <EricDalquist> but you're still a whole version behind
[16:00:37 EST(-0500)] <EricDalquist> and yeah
[16:00:38 EST(-0500)] <EricDalquist> that too
[16:01:03 EST(-0500)] <agherna> so we might be running in to a "middleware" issue then...
[16:01:05 EST(-0500)] <athena7> actually, we got bitten a few months ago by using the old classes12.jar with up2.6
[16:01:15 EST(-0500)] <EricDalquist> so go get a ojdbc14.jar for your database version (or just get the latest since the drivers are backwards compatible)
[16:01:16 EST(-0500)] <athena7> i don't think we had the same problem you're seeing
[16:01:22 EST(-0500)] <athena7> but i definitely would encourage upgrading
[16:01:26 EST(-0500)] <EricDalquist> we aggresively upgrade our ojdbc driver
[16:01:40 EST(-0500)] <agherna> good advice
[16:01:42 EST(-0500)] <athena7> we had mysterious, occasional database corruption that turned out to be a result of using the old driver
[16:01:42 EST(-0500)] <EricDalquist> there is such a track record of bugs in it that any fixes are helpful
[16:01:49 EST(-0500)] * EiNZTEiN (n=einztein@205.241.143.4) has joined ##uportal
[16:01:55 EST(-0500)] <athena7> it periodically ate XML that was stored as CLOBs
[16:02:06 EST(-0500)] <agherna> ouch
[16:02:14 EST(-0500)] <dstn> they completely got rid of the oracle.jdbc.driver package in the latest...heads up on that
[16:02:39 EST(-0500)] <dstn> its a deprecated package but if you have anything that uses it it won't work with the newest
[16:02:53 EST(-0500)] <agherna> so we would need to change to oracle.jdbc.OracleDriver, right>
[16:02:57 EST(-0500)] <dstn> ya
[16:02:58 EST(-0500)] <agherna> er ? not >
[16:03:16 EST(-0500)] <agherna> this is all good to know
[16:03:31 EST(-0500)] <agherna> it could explain some of what we're seeing
[16:03:38 EST(-0500)] <EricDalquist> no
[16:03:45 EST(-0500)] <agherna> ?
[16:03:47 EST(-0500)] <dstn> that's only if your using ojdbc6.jar
[16:03:59 EST(-0500)] <EricDalquist> oracle.jdbc.OracleDriver is not in the oracle.jdbc.driver package ...
[16:04:05 EST(-0500)] <dstn> anything before that it will still have the package, although its deprecated
[16:04:22 EST(-0500)] <EricDalquist> and you will only see issues if you have code that uses the Oracle APIs directly
[16:04:33 EST(-0500)] <agherna> which we don't...
[16:04:38 EST(-0500)] <EricDalquist> and ojdbc6 I'm assuming is for JDK6+
[16:04:48 EST(-0500)] <EricDalquist> which uses JDBC3
[16:04:49 EST(-0500)] <dstn> and 11
[16:04:51 EST(-0500)] <dstn> 11g*
[16:04:56 EST(-0500)] <agherna> we run 1.5.0_something
[16:05:05 EST(-0500)] <agherna> jdk1.5.0_something that is
[16:05:09 EST(-0500)] <agherna> but anyway
[16:06:09 EST(-0500)] <agherna> what I'm reading from you guys is that classes12.jar should be replaced with ojdbc14.jar and our rdbm.properties file needs to have the driver class changed to account for that (to oracle.jdbc.OracleDriver).
[16:06:37 EST(-0500)] <agherna> some of you have also had mysterious things happen that could not be explained when using the drivers in classes12.jar
[16:06:46 EST(-0500)] <EricDalquist> um
[16:06:57 EST(-0500)] <EricDalquist> I just copied that class name from your post
[16:06:58 EST(-0500)] <EiNZTEiN> wow
[16:07:00 EST(-0500)] <EricDalquist> we're using oracle.jdbc.driver.OracleDriver
[16:07:03 EST(-0500)] <EiNZTEiN> uP3 looks a lot better (wink)
[16:07:11 EST(-0500)] <EricDalquist> (smile)
[16:07:14 EST(-0500)] <EiNZTEiN> whaat is the default user/pass now? (smile)
[16:07:22 EST(-0500)] <EricDalquist> same as it was in 2
[16:07:27 EST(-0500)] <EiNZTEiN> admin/admin? :|
[16:07:35 EST(-0500)] <EricDalquist> admin/admin, demo/demo, student/student, developer/developer
[16:07:37 EST(-0500)] <EiNZTEiN> because it's not working
[16:07:38 EST(-0500)] <EiNZTEiN> :S
[16:07:40 EST(-0500)] <dstn> You don't have to switch to oracle.jdbc package...although its recommended since it was deprecated in 9.
[16:07:53 EST(-0500)] <agherna> OK
[16:08:13 EST(-0500)] <agherna> but buggy driver code could help explain what we're seeing
[16:08:28 EST(-0500)] <EricDalquist> the cas login is failing?
[16:08:43 EST(-0500)] <agherna> the "maximum number open cursors exceeded" has bitten us 3 times already
[16:08:45 EST(-0500)] <dstn> "In every release since 9.0.1 we have encouraged customers to switch to oracle.jdbc and stated that oracle.jdbc.driver would be desupported." <-- that's from the readme
[16:09:10 EST(-0500)] <agherna> and our dbas have set us up with 1000 open cursors
[16:09:52 EST(-0500)] <EiNZTEiN> ohh
[16:09:54 EST(-0500)] <EiNZTEiN> I see the problem
[16:10:03 EST(-0500)] <EiNZTEiN> is there anyway I can change the "localhost"?
[16:10:11 EST(-0500)] <EricDalquist> you have to edit web.xml and security.properties
[16:10:13 EST(-0500)] <EiNZTEiN> cuz I'm running the quick-start on a dev machine next to me
[16:10:14 EST(-0500)] <EiNZTEiN> ok ok
[16:10:17 EST(-0500)] <EiNZTEiN> thanks
[16:10:19 EST(-0500)] <EricDalquist> yup
[16:10:42 EST(-0500)] <agherna> which we potentially could exceed--we have about 12000 potential users that will be hitting this thing and we've had just 150 users bring it down...
[16:10:47 EST(-0500)] <agherna> which is weird
[16:11:25 EST(-0500)] <agherna> but if the middleware is buggy and the uPortal code is closing its statements correctly then it has to be the middleware (somewhere)
[16:11:56 EST(-0500)] <dstn> Ya that's not good.
[16:12:09 EST(-0500)] <agherna> yeah, especially since it can't be debugged
[16:12:12 EST(-0500)] <agherna> well,
[16:12:49 EST(-0500)] <agherna> you can debug it. the road to getting there will have Oracle's lawyers at your door if they caught wind of it though.
[16:13:01 EST(-0500)] <EricDalquist> (smile)
[16:14:04 EST(-0500)] <jayshao> can you trace the db on the oracle side?
[16:14:11 EST(-0500)] <jayshao> see what statements are holding the cursors open?
[16:14:19 EST(-0500)] <agherna> yes, we've done some of that
[16:14:35 EST(-0500)] <jayshao> anything useful?
[16:14:42 EST(-0500)] <agherna> we have our dbas taking a snapshot of the database with the open cursors and the SQL used to open them
[16:14:43 EST(-0500)] <EiNZTEiN> EricDalquist, I don't see where to change any host related info on those 2 files
[16:14:44 EST(-0500)] <EiNZTEiN> :S
[16:14:51 EST(-0500)] <agherna> so far, not that we can really tell.
[16:15:07 EST(-0500)] <EricDalquist> in uPortal's web.xml there is a CASValidateFilter
[16:15:13 EST(-0500)] <EricDalquist> that has some init params
[16:15:23 EST(-0500)] <EricDalquist> in security.properties you need to change the login/logout URLs
[16:15:54 EST(-0500)] <agherna> we can locate the sql in the code
[16:18:20 EST(-0500)] <EiNZTEiN> EricDalquist, no "CASValidateFilter" anywhere
[16:18:20 EST(-0500)] <EiNZTEiN> :S
[16:19:09 EST(-0500)] <EricDalquist> https://www.ja-sig.org/svn/up2/trunk/uportal-war/src/main/webapp/WEB-INF/web.xml
[16:20:09 EST(-0500)] <EricDalquist> about 1/3 of the way down
[16:21:05 EST(-0500)] * colinclark (n=colin@srb-wlan-35.AirBears.Berkeley.EDU) has joined ##uportal
[16:22:32 EST(-0500)] <EricDalquist> find it?
[16:22:56 EST(-0500)] <EricDalquist> in the quickstart it would be in uPortal-3.0.0-RC2/uportal-war/src/main/webapp/WEB-INF/web.xml
[16:23:10 EST(-0500)] <EricDalquist> then you would need to run ant deploy-war after making the change
[16:27:36 EST(-0500)] * tcarroll (n=tcarroll@flatlander.ci.uiuc.edu) has joined ##uportal
[16:28:31 EST(-0500)] <EiNZTEiN> checking
[16:30:13 EST(-0500)] <EiNZTEiN> I don't have that whole section :|
[16:30:23 EST(-0500)] * tcarroll (n=tcarroll@flatlander.ci.uiuc.edu) has left ##uportal
[16:31:05 EST(-0500)] <EiNZTEiN> here's my web.xml
[16:31:06 EST(-0500)] <EiNZTEiN> http://pastie.caboo.se/162509
[16:32:53 EST(-0500)] <EricDalquist> this is from the RC2 quickstart download on uportal.org?
[16:32:58 EST(-0500)] <EiNZTEiN> yup
[16:33:01 EST(-0500)] <EiNZTEiN> :S
[16:33:05 EST(-0500)] <EricDalquist> uhoh
[16:33:12 EST(-0500)] <EiNZTEiN> (sad)
[16:33:23 EST(-0500)] <EiNZTEiN> tar.gz distrib (if it makes any dif)
[16:33:35 EST(-0500)] <EricDalquist> ok
[16:33:41 EST(-0500)] <EricDalquist> quick start or dev quick start?
[16:33:51 EST(-0500)] <EiNZTEiN> hmm
[16:33:53 EST(-0500)] <EiNZTEiN> I'll check
[16:34:12 EST(-0500)] <EiNZTEiN> uPortal-3.0.0-RC2-quick-start.tar.gz
[16:34:23 EST(-0500)] <EiNZTEiN> I guess non-dev
[16:34:37 EST(-0500)] <EricDalquist> ok
[16:37:35 EST(-0500)] <EricDalquist> I just downloaded http://www.ja-sig.org/downloads/uportal/uPortal-3.0.0-RC2/uPortal-3.0.0-RC2-quick-start.tar.gz
[16:37:44 EST(-0500)] <EricDalquist> and the included web.xml has the cas filter section in it
[16:37:58 EST(-0500)] <EricDalquist> do you have more that one quickstart on that box?
[16:47:42 EST(-0500)] * colinclark (n=colin@srb-wlan-5.AirBears.Berkeley.EDU) has joined ##uportal
[16:58:16 EST(-0500)] <EiNZTEiN> nop
[16:58:19 EST(-0500)] <EiNZTEiN> weirddd
[16:58:21 EST(-0500)] <EiNZTEiN> :|
[17:00:12 EST(-0500)] <EricDalquist> somethings not right or out of date then
[17:03:32 EST(-0500)] <EiNZTEiN> I downloaded like 1h ago :|
[17:03:37 EST(-0500)] <EiNZTEiN> I'll re-download
[17:18:46 EST(-0500)] * awills (n=awills@12.164.136.183) has joined ##uportal
[17:19:31 EST(-0500)] <awills> by chance is anyone here familiar with running m2 from within ant?
[17:19:40 EST(-0500)] <EricDalquist> yup
[17:19:43 EST(-0500)] <EricDalquist> up3 does it lots
[17:19:52 EST(-0500)] <awills> the macrodef stuff?
[17:19:57 EST(-0500)] <EricDalquist> yup
[17:20:04 EST(-0500)] <EricDalquist> macrodefs + the maven ant jar
[17:20:16 EST(-0500)] <awills> hmm... yeah i suppose that would have to be the case
[17:20:31 EST(-0500)] <awills> since uP3 still supports > ant initportal
[17:20:48 EST(-0500)] <EricDalquist> the macrodefs are smart enough to do conditional recompiles and such
[17:21:18 EST(-0500)] <EricDalquist> are you working with a multi-module maven project?
[17:21:35 EST(-0500)] * theclown (n=clown@142.150.154.101) has left ##uportal
[17:21:51 EST(-0500)] <awills> yes
[17:22:49 EST(-0500)] <awills> since the <macrodef> use the maven ant jar, does that mean mvn doesn't need to be installed on the server itself?
[17:23:26 EST(-0500)] <EricDalquist> no it does
[17:23:43 EST(-0500)] <EricDalquist> if you look at the uportal build it calls the mvn command line
[17:23:56 EST(-0500)] <EricDalquist> I have to catch the bus home but we can chat more about this tomorrow
[17:23:59 EST(-0500)] <awills> ah bummer
[17:24:07 EST(-0500)] <awills> sure
[17:24:33 EST(-0500)] <EricDalquist> http://erics-notes.blogspot.com/2008/01/using-ant-with-maven-project.html
[17:24:51 EST(-0500)] <EricDalquist> there is a way to use ant to front a maven project (single module) on a machine without maven support
[17:24:54 EST(-0500)] <EricDalquist> later
[17:41:51 EST(-0500)] <EiNZTEiN> hmmm
[17:41:56 EST(-0500)] <EiNZTEiN> "ant deploy-war" doesn't exist
[17:41:57 EST(-0500)] <EiNZTEiN> :S
[17:45:38 EST(-0500)] <EiNZTEiN> oops
[17:45:40 EST(-0500)] <EiNZTEiN> go it (wink)
[17:45:41 EST(-0500)] <EiNZTEiN> wrong dir
[17:45:42 EST(-0500)] <EiNZTEiN> hehe

  • No labels