> with HZ=1000, that's almost 25 days. i hear today about the being > something fixed with HZ=4096, which is only 6 days.
I once worked on something where HZ=8000 was needed. > i recommend avoiding any kernel timeouts longer than a day. With HZ=8000, tick counts go negative after about 3.1 days. "avoiding...longer than a day" is still safe, but has uncomfortably small margin. It might be wise to limit it based on ticks, rather than based on time-to-sleep, using multiple successive timeouts if the wait time is longer than (say) 1<<30 ticks.... /~\ The ASCII Mouse \ / Ribbon Campaign X Against HTML mo...@rodents-montreal.org / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
