The problem is the same as the original post: something bad happens, swap gets used or over-used, and the machine locks. Without even a warning message. Linux always behaved that way. BSD-derived OS's running on the very same commodity Intel hardware dont have that problem. Among my fellow system admins the rule-of-thumb became "Don't swap". Give it enough RAM to prevent that or re-distribute application load to prevent it. If you cant afford that, well.... Why does the linux kernel lock the machine without messaging when it experiences virtual memory pressure?
Dan Ritter direct reply to your email addr bounces. On Fri, Nov 13, 2020, 9:20 AM Dan Ritter <d...@randomstring.org> wrote: > Nicholas Geovanis wrote: > > I guess Im not the only crank with antique hardware. One of my few > unending > > beefs with the linux kernel is swap behavior. Everyone knows what it's > for > > and how it "works". But even glancing thru the code doesn't explain its > > real-time run-time behavior. In contrast, the last time I had swap issues > > like that on a BSD-line unix OS was 35 years ago (on DEC hardware ;-) > Same > > thing with commercial Solaris, HP/UX, AIX. What is the linux kernel doing > > wrong? > > > > Want to start a new thread and explain what the problem is? > > -dsr- >
