I have a master server running Solaris 8 and Veritas 3.4, and I have a
group of Windows 2000 systems I've just added to the client lists, and
when I try to back them up, they run for as little as 10 minutes (or
all the way up to an hour or more) but eventually I get the following
error:

15:03:55 [2095] <4> bpbrm main: from client xxx: INF
-                      OTM: enabled
15:09:40 [2095] <32> bpbrm main: from client xxx: FTL - tar file write
error (10054)
15:09:40 [2095] <4> bpbrm main: client xxx EXIT STATUS = 14: file
write failed
15:12:11 [2095] <2> sighdl: pipe signal
15:12:11 [2095] <16> put_long: (11) network write() error: Broken pipe
(32); socket = 4
15:12:11 [2095] <16> inform_client_of_status: could not send server
status message
15:12:11 [2095] <8> inform_client_of_status: Could not set linger
value on socket. Errno = 22: Invalid argument
15:12:11 [2095] <2> getsockconnected: host=xxx service=bpcd
address=x.x.x.x protocol=tcp reserved port=13782
15:12:11 [2095] <2> getsockconnected: Connect to xxx on port 735
15:12:14 [2095] <2> bpcr_get_version_rqst: bpcd version: 03400000
15:12:14 [2095] <2> bpcr_get_version_rqst: bpcd version: 03400000
15:12:14 [2095] <2> bpcr_get_version_rqst: bpcd version: 03400000
15:12:14 [2095] <4> bpbrm Exit: client backup EXIT STATUS 14: file
write failed

I've checked the NIC card and the switch configurations, and that
doesn't seem to be it. I have other systems with the same OS with no
problem. I have also tried upping the timeout on the client end, with
no added success. Googling doesn't seem to yeild any useful info. I
have three systems (on the same subnet and switch) exhibiting this
behavior. 

Any and all suggestions gratefully appreciated...

-Suzanne