Here's a quick analysis of the tcpdump that Dave Dykstra just sent out. There seems to be a couple of (possibly unrelated) problems. On line 2066, static acks 69041, so on line 2067, dynamic sends out 69041:69628. Static should ack this, but it never does, so dynamic just keeps retrying that block and never getting an ack. (This may be related to static being out of receive window space, but dynamic seems to be following the rules. On line 2061, static gave it permission to send 1460 bytes, so dynamic did so on lines 2064 and 2065.) Then, on line 2092, static sent some old data (1583721:1585181) which was already acked back on 2085. It seems to keep sending this block over and over again, even though dynamic is acking up to 1592481. So, it's looping at that point - dynamic keeps sending 69041:70501 over and over again because it only gets an ack 69041, and static keeps sending 1583721:1585181 and getting an ack for 1592481. I don't have a very deep understanding of TCP, but both of these look like problems on the Linux side to me. Pete Fales
Re: I also am getting hang/timeout using rsync 2.4.6 --daemon
Lucent Central Exptools Administrator Wed, 25 Oct 2000 02:36:31 -0700
- Re: I also am getting hang/timeout u... Neil Schellenberger
- Re: I also am getting hang/time... Dave Dykstra
- Re: I also am getting hang/... Lucent Central Exptools Administrator
- Re: I also am getting hang/... Dave Dykstra
- Re: I also am getting hang/time... Dave Dykstra
- Re: I also am getting hang/... Dave Dykstra
- Re: I also am getting hang/time... Dave Dykstra
- Re: I also am getting hang/time... Dave Dykstra
- Re: I also am getting hang/time... Scott Russell
- Re: I also am getting hang/time... Eric Whiting
- Re: I also am getting hang/time... Dave Dykstra
- Re: I also am getting hang/time... Scott Russell