Comment 42 for bug 1655842

Don't forget that the earlier kernels are affected by Bug #1647400, which does something even worse (hang the system). I've verified that it affected my particular system before 4.4.0-59, and it may explain a couple of lockups I had previously experienced during normal operation when using previous kernels. -59 fixes the bug but introduces the permature OOM kill issue; if it weren't for the kernels currently in proposed (assuming they indeed fix this bug), I wouldn't really have a reliable kernel at all to use.

With the 4.4.0-59 kernel, I got hit with two unexplained OOM kills, each occurring within about 3 days of uptime. I then tested the -62 kernel in proposed for just under 14 days and didn't see any OOM kills, and I've now been testing -63 for a couple of days and haven't any issues yet. However, it might help if anyone has an idea how the OOM kill bug might be reliably reproduced. “5 working days” isn't very long to reliably be sure the problem is solved otherwise; it took more than half that time upon upgrading to -59 for me to hit the bug by chance.