I'm having an odd problem with Taverna 1.7.2 on Mac OS X 10.5. (The workflow uses JDBC, so Taverna 2 is not an option.)
I am trying to run a workflow multiple times to process different batches of data (a total of about 1000 runs of the workflow). To automate the process, I wrote a simple perl script that runs through the various parameters for the workflow, puts the parameters into files and calls the "no GUI" shell script to run the workflow with the inputs taken from the files. (In passing, it would be nice to be able to pass workflow inputs as shell arguments when they're simple things like one number.) BTW, I originally tried doing this all in Taverna by having a single workflow that invoked an embedded workflow repeatedly but that ran slower and slower and slower until it just stopped after a few hundred iterations. I suspect that some part of the workflow is creating objects that aren't getting freed and so Taverna is running out of memory, but I didn't investigate in detail. Hence the new approach of running the workflow as a fresh instance of Taverna for each set of parameters. (Actually, it's a bit more complicated, but that's the basic idea.) The problem I've run into is that after a few dozen executions of Taverna, the workflow starts to fail with an odd message: Shared archive: sharing disabled for server vm There is a problem connecting to the url:http://www.mygrid.org.uk/taverna/updates/1.7.2/taverna-1.7.2-profiles.xml There is a problem connecting to the url:http://bioinf.ncl.ac.uk/mirror/taverna/updates/1.7.2/taverna-1.7.2-profiles.xml There is a problem connecting to the url:http://mirror.omii.ac.uk/taverna/updates/1.7.2/taverna-1.7.2-profiles.xml Unable to connect to any of the mirror sites for to check for updated profiles. THIS STRONGLY INDICATES YOU HAVE NO NETWORK ACCESS. (Followed by errors accessing the various web services and the workflow failing.) The first time this happened, I figured that perhaps the server had fallen off the network momentarily, and restarted the perl script, which worked. However, it's happening repeatedly and as I can ssh into the server to restart the run, it seems that the server does have network access; it's just Taverna that is having trouble for some reason. Indeed, I can be logged into the server watching the run and get that message. Restarting the perl script usually gets things working again, at least for another few dozen runs. Since restarting the run seemed to work, I modified the perl script to check if the workflow had finished by checking for the output file, and if not, to wait 5 minutes and rerun the workflow. Interestingly, that does not work--the attempt to rerun the workflow fails with the same message. But if I kill the perl script and restart it, the next run of the same data almost always does work. I'm not sure what the difference is, but apparently there is one. Since my attempts to work around the apparent network problem has been unsuccessful, I'm wondering if I can solve them. Any ideas about why Taverna might suddenly not be able to access the network? Or other ideas for a different work around? Or even solutions to getting it all to work in Taverna? Kevin Crowston Syracuse University Phone: +1 (315) 464-0272 School of Information Studies Fax: +1 (866) 265-7407 348 Hinds Hall Web: http://crowston.syr.edu/ Syracuse, NY 13244-4100 USA ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ taverna-users mailing list [email protected] [email protected] Web site: http://www.taverna.org.uk Mailing lists: http://www.taverna.org.uk/taverna-mailing-lists/
