Greetings everyone,

Yesterday we encountered pretty much 100% errors on all workunits sent out by the grid. These errors exited immediately due to certain input files not found. This was caused by a new backend server script to help improve performance and speed of loading new work. However, I forgot to add in the clause that would tell certain work units to copy_file to the slots directory on your machine. The science application did not know how to handle BOINC symbolic links and came back with errors. I have corrected the issue with the load script, however right now I'm working to correct the workunit definitions in the database so that workunits sent out going forward work properly.

Side note: If you were wondering about the performance gains of the new script, it is between 3-10 times faster depending on the project, which also allows us to keep the 48 hour buffer of work on the grid caught up by running the new script for about a minute every 30 minutes.

Again, thank you for your participation and patience as we work through this issue.

Thanks,
-Uplinger