On Tue, Mar 09, 2010 at 11:27:56AM -0800, Matthew Dillon wrote: > fetch http://apollo.backplane.com/DFlyMisc/hammer12.patch > > Hopefully it will print out the debugging message and continue on its > merry way instead of asserting in the higher level HAMMER code. > > What I'm trying to do is to catch the one unlocked cursor adjustment > which I believe can cause the problem by setting a flag, and then > adjusting the iteration code to skip any records which might have > wiggled their way in. Hopefully then the higher level code will > not assert on the obj_id being incorrect.
I'm running pbulk on a few machines around me with the patch applied. I'll let you know if anything unusual happens again. Thanks.
