15.04.2020 09:48, David C. Rankin пишет:
All,
I think it was James that had a similar question, but at the time I had no context to compare it to. But I had an occurrence last night where running valgrind on a simple bit of code that allocated 4.5G causes the my box to swap.
The machine was in hardlock during swap -- no mouse pointer movement, not keyboard control -- no nothing. I was contemplating a hard reset while I sat waiting... and waiting... and waiting (probably no more than 3 minutes -- but that provides a lot of time to think about whether you need to nuke it or not)
Try increasing /proc/sys/vm/dirty_ratio to very large value, e.g. 95. Do you still observe these stalls? Do not leave it that that in normal case.
I have had boxes swap hundreds and hundreds of times over the decades, but never had a box exhibit this "hard-lock" behavior before. I now valgrid sets up its own protected environment to do its work in, but I don't see why it swapping would have any worse of different impact on the system than anything else swapping (I just don't know -- but wouldn't expect it to)
Is this a bug, a feature, a valgrind special environment issue? If I recall the earlier thread (there have been a couple - Why swap?, etc.. recently), James or whoever it was was seeing similar unresponsiveness during the swap.
I ran into this answering a question on StackOverflow. With 8G of RAM total allocating 4.5 while you have a full desktop with a dozen or so tabs in Firefox, etc. will cause your box to swap. Simple example code posted at:
https://paste.opensuse.org/3916734
Why would the system appear locked while valgrind was swapping?
-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org