[bind10-dev] failed lettuce master 812392465ae1ffa4a583186fd48f4c7aff1640dd

[JINMEI Tatuya / 神明達哉]

At Wed, 24 Oct 2012 12:29:20 +0200,
Shane Kerr <shane at isc.org> wrote:

> > > 2012-10-22 10:26:27.674 INFO  [b10-boss.boss] BIND10_PROCESS_ENDED 
> > > process 79181 of b10-auth ended with status 138 2012-10-22
> > > 10:26:27.674 ERROR [b10-boss.boss] BIND10_COMPONENT_FAILED
> > > component b10-auth (pid 79181) failed: process dumped core with
> > > exit status 138 (killed by signal 10: SIGBUS)
> > > 
> > > Here is the backtrace of /cores/core.79181
> > 
> > I couldn't reproduce this for the latest master
> > (3404981c39241b4463e661981614ec0d157b924d) on the macmini box.
> 
> That's scary. Is there anything that can cause SIGBUS on OS X other than
> a coding error? (Like... broken fan making the box overheat, or bad
> memory, or...)

The version that caused the SIGBUG seems to be before the (second)
revert of #2198 (see http://bind10.isc.org/ticket/2198#comment:28).
Since it caused various race conditions, it's not surprising if it's
the reason for this type of strange failures.  Since the revert the
lettuce tests don't seem to fail, so I think we can forget it at least
for now.

---
JINMEI, Tatuya