iz (iz.delete@this.iz.com) on March 16, 2021 8:36 am wrote:
> Andrey (andrey.semashev.delete@this.gmail.com) on March 16, 2021 2:54 am wrote:
> > If the software crashes in some conditions, there is no reason to believe it'll
> > stop crashing after a restart. Restart is not a fix for a bug in code.
>
> That is mostly a problem if you restore state and try to continue without interruption.
> I was talking about restarting in general. Of course you can't expect programs
> to be able to restart with full state restored without much effort.

It's not just about restoring previous state and continuing from there. Most workloads involve repetitive activities, like processing many more-or-less similar user requests or processing similar data bundles or whatever. If a piece of software crashed on one iteration then it is very likely going to crash after restart on the next similar activity.

There are, of course, cases when the software crashes every once in a while, when Moon is in the right position. Restarts do help to keep the system mostly running when these happen. However, such cases are more rare than the other, more "stable" kind of bugs.