Alas, I only have 1 datapoint, a 2013 MacBook Pro that was upgraded. Everything else is either still on older versions or are running Linux, Windows, etc.
The symptom: Post upgrading, this MacBook will now decide that hosts on the WiFi LAN are unreachable. Established connections will be terminated (notably SSH), and pings are immediately rejected locally with ping: sendto: Host is down. This usually means the networking stack can't resolve the MAC address of the remote host any longer, and, indeed, running arp shows that the host of interest has an at (incomplete) on en0 ifscope [ethernet] entry.
Of course, no other machines are having these types of connectivity issues, and the target systems in question are supporting other connections happily.
This sounds vaguely like an old ARP bug (net.link.ether.inet.arp_unicast_lim=0) rearing its head again, though the symptoms (double-NAT, etc.) don't quite align. I'll see if setting that helps at all, but I'm curious if other network admins are hearing similar complaints yet.
No comments:
Post a Comment