> According to the outputs, e1000g has already passed > the receiving > packets up. > At some point the packets are dropped in the stack. > Please find the ip > and icmp statistics. > kstat -m ip -n ip 1 > kstat -m ip -n icmp 1
This morning the Solaris system was still up. I woke my MacBook from sleep, copied a file to the ZFS share, copied that file from the ZFS share to a Windows machine, and the Solaris server hung. Not sure if that's correlation or causation, but I should also note that for some reason my Windows machine can't mount a CIFS share if the Mac is already sharing (but not vice-versa). This bug might be triggered by some sort of race condition. Anyway, here's this morning's test. Pings are being sent from Solaris to the Mac, Wireshark running on the mac is reporting that the pings are being received and replied to, but the replies aren't being heard. By the way I've got a sysadmin friend working for eBay (a major Solaris customer) who told me this: "so apparently we've seen a bug in the last week, where some machines will unregister their mac addresses. i wonder if it's related. i will let you know what i hear back. we have platinum tech support, so i'll probably hear sooner than you will." Could that be what's happening? Can Solaris send just fine, but it's not getting the replies (and sending out a lot of ARP "who has my IP" packets) because it's lost its own mac address so it doesn't know what to listen for anymore? Thanks again for your continued help. -- This message posted from opensolaris.org _______________________________________________ networking-discuss mailing list networking-discuss@opensolaris.org
