From legnitto at ucsc.edu Thu May 8 16:21:49 2003 From: legnitto at ucsc.edu (Christian Legnitto) Date: Thu, 08 May 2003 16:21:49 -0700 Subject: Problems with Monarch (possible bug?) Message-ID: <3EBAE68D.6030804@ucsc.edu> Hey all, Christian Legnitto here in Santa Cruz. Rex is having me setup monarch and metacat here at UCSC. I am having very odd issues and was wondering if anyone has seen/fixed the same problems. Version Information: ------------------------------------------------------- Metacat Version: 1.2.1 * Checked out from CVS on May 8th using 'cvs co -r CHANGES_FOR_MONARCH_01062003 metacat' Monarch Version: 0.0.0 Alpha1 * Checked out from CVS on May 6th using 'cvs co monarch' Java Version: Sun 1.4.1_02 *Using standard pipelines that came with Monarch Everything built properly with no errors. ------------------------------------------------------- I have two problems actually... 1) I access monarch through metacat (http:///catalog/monarch-index.html), select a dataset (via checkbox), then select the desired pipeline (via radio button). At the next screen (selecting databindings) the selection boxes are not populated with choices. You can not run pipelines because they do not know what the inputs are if you cannot select them at the selection page. The error happens every time. Restarting tomcat, rebuilding monarch/metacat, and reloading the page does not fix anything. It can be fixed though, but it requires a little voodoo. On the main menu I have to click on the graphIcon.jpg (which runs the simple scatterplot). Then I get an internal server error: javax.servlet.ServletException: edu.ucsb.nceas.monarch.MonarchMetacatInterfaceServlet Error processing servlet action. Server returned HTTP response code: 500 for URL: http:///monarch/servlet/monarch java.io.IOException: Server returned HTTP response code: 500 for URL: http:///monarch/servlet/monarch (which I assume is monarch dying on the pipeline because it could not get the inputs) Then if I go back and select a dataset and the desired pipeline again the selection boxes are magically populated. The selection boxes will be populated for a bit, but if I leave monarch alone for an hour and try it again they will not be populated again. This leads me to believe it is a cache problem. My data directory has the right permissions and the data is in there. The cache directory has the right permissions as well. The data is NOT in the cache until I click on the icon (but it *should* be in the cache when I run a pipeline...it isn't) This is a really weird error, and I do not know if it is metacat, monarch, or my setup. It could also be incompatible versions of metacat/monarch. Anyone have any ideas? Then the next problem..... 2) When the selection boxes ARE populated, if I try to run a pipeline SAS terminates with errors. The error files have a bunch of ^M's all over the place. SAS has errors on the lines with ^M's. Both the code in the steps and the generated runfile.sas have no ^M's in them (looked with vi), so I cannot figure out where they are generated. SAS works fine with other programs. So far setting up metacat and monarch has been trivial, but getting them to work together correctly has not. Does anyone have a working Metacat+Monarch install right now? If so, which sources were they built from? Could you send them to me :^) ? Thanks for listening to my longwinded plea for help :^) - Christian Legnitto From berkley at nceas.ucsb.edu Sat May 10 11:18:00 2003 From: berkley at nceas.ucsb.edu (Chad Berkley) Date: Sat, 10 May 2003 11:18:00 -0700 Subject: Problems with Monarch (possible bug?) In-Reply-To: <3EBAE68D.6030804@ucsc.edu> References: <3EBAE68D.6030804@ucsc.edu> Message-ID: <1052590680.3ebd4258d5430@hyperion.nceas.ucsb.edu> Hey Christian, It sounds like you may have a windows/linux compatibility problem. Are you running monarch and metacat on the same machine? Is it a linux box or windows? What version of java and tomcat do you have? You may have noticed that at the top of monarch's build.xml file, there are two targets aptly named "rich" and "chad". They should actually be called "windows" and "linux" respectively. You can choose which one intializes by changing the depends on the init target. I have metacat and monarch working together on my machine. I use the head of CVS for metacat though. I haven't tested monarch with the release 1.2.1, though I probably should. we had some early incompatibilities with the apache jar files in metacat and monarch. this may also be a problem. If you want to be a guinea pig for a new product, you could check out the cvs module called "overlay" and try to use it. It is an attempt to get metacat and monarch to install together in a simple way. It works on my linux box at least so you might try it out. The build file is pretty self explanatory. If you need any help interpreting it let me know. I'm at a meeting all this week so I won't be able to do much for you until I get back to nceas, but if you want to chat, I'm on IRC. I think the most major thing you can do to try to fix your problem(s) is to upgrade metacat to the head of CVS. chad Chad Berkley Metadata Systems Developer National Center for Ecological Analysis and Synthesis 735 State St. Suite 300 Santa Barbara, CA 93117 Quoting Christian Legnitto : > Hey all, Christian Legnitto here in Santa Cruz. Rex is having me setup > monarch and metacat here at UCSC. I am having very odd issues and was > wondering if anyone has seen/fixed the same problems. > > Version Information: > ------------------------------------------------------- > Metacat Version: 1.2.1 > * Checked out from CVS on May 8th using > 'cvs co -r CHANGES_FOR_MONARCH_01062003 metacat' > Monarch Version: 0.0.0 Alpha1 > * Checked out from CVS on May 6th using > 'cvs co monarch' > Java Version: Sun 1.4.1_02 > > *Using standard pipelines that came with Monarch > > Everything built properly with no errors. > ------------------------------------------------------- > > I have two problems actually... > > 1) I access monarch through metacat > (http:///catalog/monarch-index.html), select a dataset (via > checkbox), then select the desired pipeline (via radio button). At the > next screen (selecting databindings) the selection boxes are not > populated with choices. You can not run pipelines because they do not > know what the inputs are if you cannot select them at the selection > page. The error happens every time. Restarting tomcat, rebuilding > monarch/metacat, and reloading the page does not fix anything. > > It can be fixed though, but it requires a little voodoo. On the main > menu I have to click on the graphIcon.jpg (which runs the simple > scatterplot). Then I get an internal server error: > > javax.servlet.ServletException: > edu.ucsb.nceas.monarch.MonarchMetacatInterfaceServlet > Error processing servlet action. > Server returned HTTP response code: 500 for URL: > http:///monarch/servlet/monarch > java.io.IOException: Server returned HTTP response code: 500 for URL: > http:///monarch/servlet/monarch > > (which I assume is monarch dying on the pipeline because it could not > get the inputs) > > Then if I go back and select a dataset and the desired pipeline again > the selection boxes are magically populated. > > The selection boxes will be populated for a bit, but if I leave monarch > alone for an hour and try it again they will not be populated again. > > This leads me to believe it is a cache problem. My data directory has > the right permissions and the data is in there. The cache directory has > the right permissions as well. The data is NOT in the cache until I > click on the icon (but it *should* be in the cache when I run a > pipeline...it isn't) > > This is a really weird error, and I do not know if it is metacat, > monarch, or my setup. It could also be incompatible versions of > metacat/monarch. Anyone have any ideas? > > Then the next problem..... > > 2) When the selection boxes ARE populated, if I try to run a pipeline > SAS terminates with errors. The error files have a bunch of ^M's all > over the place. SAS has errors on the lines with ^M's. Both the code in > the steps and the generated runfile.sas have no ^M's in them (looked > with vi), so I cannot figure out where they are generated. SAS works > fine with other programs. > > > So far setting up metacat and monarch has been trivial, but getting them > to work together correctly has not. Does anyone have a working > Metacat+Monarch install right now? If so, which sources were they built > from? Could you send them to me :^) ? > > Thanks for listening to my longwinded plea for help :^) > > - Christian Legnitto > > _______________________________________________ > monarch-dev mailing list > monarch-dev at ecoinformatics.org > http://www.ecoinformatics.org/mailman/listinfo/monarch-dev > ------------------------------------------------- This mail sent through IMP: http://horde.org/imp/