The last one I'm aware of that did not exhibit this issue was 3.5.0-27. I wish I had a simpler repro though, since on our system it takes 10-15 hours of heavy processing to hit the uninterruptible sleeps.
Could it be tracked by looking at the state of the OS? Every new lxc-start ends up hanging after it happens.
The last one I'm aware of that did not exhibit this issue was 3.5.0-27. I wish I had a simpler repro though, since on our system it takes 10-15 hours of heavy processing to hit the uninterruptible sleeps.
Could it be tracked by looking at the state of the OS? Every new lxc-start ends up hanging after it happens.