So a few releases ago, I found that if I had OpenVPN running on an OpenBSD box as a hub, and I did a large transfer from one client to another, the OpenBSD box would occasionally kernel panic - something about mbufs, I can pull the kernel stack traces up if desired. The hosting company said they saw this quite often with OpenBSD boxes.
As of 4.8, I notice that if I do a similar large transfer, the system just stops responding to all network traffic for a period of time, perhaps around 10 minutes or longer. During this time it's unpingable and won't forward any packets. It eventually recovers, which is WAY better than a kernel panic (the mbuf corruption actually corrupted my root disk on more than one occasion). Before I go digging into this, I was wondering if anyone else has had this experience, and how I should consider troubleshooting it. -- http://www.subspacefield.org/~travis/ "What do you call this music?" "Hadouken" "Hadouken?" "It's down right fierce" If you are a spammer, please email j...@subspacefield.org to get blacklisted. [demime 1.01d removed an attachment of type application/pgp-signature]
