Ok, so it seems lie the following is happening: qdo submits the jobs to the pool. Normally, we just check for a timeout here, and if one occurs we're done and execute M5 locally. However, the last two days we've gotten lucky, and managed to sneak in a job in the middle of the otherwise busy queue. That's ok, however our test for the the pool being busy now passes. When SCons config starts running out compilation tests however a node isn't available, and thus eventually the configuration test fails because qdo times out. When it happens we get some crazy error like the ones we've been seeing for the past few days.
The only thing that comes to mind to fix the issue is to put a more intelligent check up-front for the pool. Something perhaps to check if there are jobs waiting to execute and bailing otherwise? Ali On Apr 14, 2009, at 2:45 AM, Cron Daemon wrote: > > See /z/m5/regression/regress-2009-04-14-03:00:01 for details. > > _______________________________________________ > m5-dev mailing list > m5-dev@m5sim.org > http://m5sim.org/mailman/listinfo/m5-dev > _______________________________________________ m5-dev mailing list m5-dev@m5sim.org http://m5sim.org/mailman/listinfo/m5-dev
