On Thu, 15 May 2008 17:06:06 -0600, LaMont Jones <[EMAIL PROTECTED]> said:
> And sadly, I'm unable to reproduce the bug, which makes fixing it rather > problematic. > If it is at all possible, could you see if the bug is present in 9.4.x? > or even 9.5.0rc1 :( I can give those a try, but I need to find a way of preventing BIND from sending notifies to our peers. I don't imagine that they'll be very happy with repeated notifies from an unauthorized host. I suppose it's possible to do this by setting up a server with e.g. a linux-vserver kernel and no network bridge. > I don't see us making much progress without some way to reproduce the > crash, or a trace of it dying... I don't know of a way to consistently reproduce the crash. To add to the fun, there are some additional problems after the OpenSSL upgrade, manifesting in a completely unresponsive nameserver, which proceeds to crash without any output in the system logs. I tried running an strace on the process to capture the moment, but the problem with this is that it not only bogs the server down a bit, it also uses up the available disk space. (Yep, with 130k+ zones, we do get a bit of traffic, and filtering in strace requires that we know what to filter for.) I'll keep on trying to reproduce the problem, though, since it does appear to happen shortly after startup, or during startup. - Jan -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
