Use of pool servers reveals unacceptable crash rate in async DNS

[Eric S. Raymond]

Kurt Roeckx <kurt at roeckx.be>:
> On Sat, Jun 25, 2016 at 11:00:39AM -0400, Eric S. Raymond wrote:
> > 
> > While this did enable me to recover from my errors, it also turned up
> > a serious problem.  The combination of the buggy async-DNS code we
> > inherited from Classic and use of pool servers causes *very* frequent
> > crashes.
> 
> Can you give some more details about the issue you're seeing?

Running under gdb, I saw a SIGBRT in the context of the async-DNS code
last night.  Because execution terminated with a signal, the near end
of the stack trace is difficult to read.  Threads are definitely
involved.

At the time I was seeing crashes every three or four minutes.
I wired force_synchronous_dns to true and the problem went away
instantly.

However I didn

> If it's the one I'm thinking about, I think the solution is to
> remove the locking of memory.

Ah yes, I think I recall that being one of the workarounds.  The
option is still available.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>