On Fri, Aug 23, 2024 at 5:09 PM Christoph Lameter (Ampere) <cl@xxxxxxxxxx> wrote: > > Look at the full kernel log output until the time of the allocation > failure? It looks like there is enough memory in every zone to satify the > allocation request. > > Stacktrace looks like memory is pushed out via kswapd and zram to > somewhere and then we get interrupted by incoming network traffic. Sure. My original email was a bit terse because it was a continuation from a discussion on IRC and I didn't explicitly say so but I already read the stacktrace. As you say, given that there's plenty of free pages in the zones including factoring in the watermarks, it's not clear to me why the allocation failed at all.
