Hi,

I've a problem with bind on gentoo : it keeps dying on me. At first I thought it was a oom related issue (Linode64), but :
- on average only 15M is allocated to processes,
- 256M of swap is now available just in case.
So I think oom isn't the issue (postgresql, postfix and tpop3d run smoothly too).

When I launch bind manually not dettaching from the terminal I see it ending on a segmentation fault. The segmentation fault doesn't occur at the same spot in the code (it fails sometimes during loading, sometimes processing a request, sometimes without having the time to trace anything, probably just after receiving a request).

The linode is on host33, it uses the 2.6.7 kernel.

I even did an "emerge --emptytree world" (sorry for the other host33 users...) and a clean reboot just in case.

Is there anybody else having bind problems on gentoo with 2.6 kernels ?
Note that the 2.6.8.1 kernel doesn't work for my linode and the 2.6.9 makes vmstat segfaulting constantly. It might very well be related to 2.6 kernels.

This *might* be related to this:
http://www.linode.com/forums/viewtopic. ... light=bind

I run Gentoo, but don't run 2.6... and I haven't run into any problems recently other than an upgrade of glibc causing problems for services still running, most notably sshd:
http://forums.gentoo.org/viewtopic.php? ... sshd+glibc

Shouldn't be nptl related as I don't have nptl in $USE (gentoo allows you to choose what you want in through the $USE var and nptl is something you can opt in or out).