>> On Sat, 6 Dec 2008 12:23:58 -0600, Roger Deschner <[EMAIL PROTECTED]> said:
> Is this perhaps a problem that could be solved another way, via > throughput thresholds? Mumble, maybe. The stall seems to happen in the middle of some critical section in TSM server land; it's really-truly in the middle of processing a TCP retransmit. I'd suspect that, if the server doesn't wake up after being in RecvW for 10 hours, that the throughput cop wouldn't hit it. In any case, It appears that the network-infrastructure oops that was making a federal case out of this has been fixed (good!) without being identified and debugged (bad!). Oh, well. Thanks! - Allen S. Rout
