Periodic (daily) Time Machine failures

rezonat0r wrote:

Many of us here (well... myself at least) are coming from Time Capsules (same network infrastructure, just using the NAS instead of TC as the backup destination). I think it's unlikely to be a network issue considering the multitude of setups in which this is occurring. Also, the size of the backup job has nothing to do with the issue. The backup fails at the authentication stage, before any backup data is transmitted. Could be something slightly different about the Time Capsule's implementation of AFP authentication as it relates to Time Machine. Or something...

I don't think you can rule out anything at this point, and we're trying to come up with a reproducible case so we have something to debug. Based on the postings, it looks like perhaps some sort of timeout could have affected the authentication. One of the logs we received had close to 15K TCP retransmits or which 130 were unrecovered, which essentially means there were unrecovered errors on the network. Whether that had anything to do with the authentication error is something I think would be useful.