From: David P G. <gr...@us...> - 2012-05-03 14:35:53
|
Steve Blackburn <Ste...@an...> wrote on 05/03/2012 06:45:18 AM: > > I'm don't have my head around this right this moment, and am under a > lot of time stress, but the following may help: > > - You're hitting a known bug. I think it may have been fixed > recently (are you running off the head?). The issue relates to the > way GC's are triggered and OOME's are thrown. The failure is > spurious (there's no reason we could not have recovered, except for > a failure in the OOME logic). The details are not in my mental > cache at this movement. My recollection is that this is something > that Daniel was looking at because it is tied in with refactoring he > did of the OOME logic. Looking through Daniel's most recent > commits may be fruitful. > I think it may be this one: http://jira.codehaus.org/browse/RVM-960. If so, it hasn't been fixed yet. Daniel made some initial investigations, but wasn't able to nail the fix in the time he had available to work on it. --dave |