Dragonfly under KVM

Gary Allan dragonfly at gallan.co.uk
Sun Jun 15 12:58:06 PDT 2008


Matthew Dillon wrote:
    I think I see what may be happening here, and I am starting to wonder
    if it is also the cause of the system lockups I am getting when testing
    HAMMER under extreme loads (with hundreds of user threads which are
    sometimes cpu-bound).
    I think it may be deadlocking between lwkt_acquire() and lwkt_schedule().
    The thread trying to migrate between cpu's is getting stuck and the 
    acquisition loop is not processing incoming IPIs while it is waiting for
    the thread to deschedule on the other cpu.

    Please try this patch.

					-Matt
					Matthew Dillon 
					<dillon at backplane.com>

Hello,

I've applied the patch and left the system running a "make -j8 
buildworld" in a loop, cleaning out /usr/obj between each run.

It has been running for over 10 hours with no lock-ups with each build 
taking ~45mins. I have also tested -j32 while taxing the disk and CPU of 
the Linux host OS.

All now appears to be Ok.

Thanks

Gary





More information about the Bugs mailing list