Here's some more detail - the brpd logs for this error on my master server.
The TCP timewait interval on the client is set to 10000, by the way.  Master
server is 120000.

-M

22:46:07.256 [29912] <2> logconnections: BPRD ACCEPT FROM
10.128.16.175.45349 TO 10.128.24.183.13720
22:46:07.257 [29912] <2> connected_peer: Connection from host
ub-23.cexp.com, 10.128.16.175, on non-reserved port 45349
22:46:07.260 [29912] <2> db_valid_master_server: ub-23.cexp.com is not a
valid server
22:46:07.260 [29912] <2> db_valid_master_server: ub-23.cexp.com is a valid
media server
22:46:07.265 [29912] <2> process_request: command C_USER_BPBACKUP_3_0 (53)
received
22:46:07.268 [29912] <2> logconnections: BPDBM CONNECT FROM
10.128.24.183.34678 TO 10.128.24.183.13721
22:46:07.954 [29912] <2> process_request: user backup request = 329199 53
oracle dba UE23 ub-23 UE23
/usr/openv/netbackup/logs/user_ops/dbext/logs/2272.0.1091767567
RMAN_Hot_ZPRD NONE 4 zprd 0 4 0 C C C C C 0 0 0 0 1
22:46:07.979 [29912] <2> getsockconnected: Connect to UE23 on port 944
22:46:07.981 [29912] <2> logconnections: BPCD CONNECT FROM 10.128.31.126.944
TO 10.128.31.107.13782
22:46:08.572 [29912] <2> logparams: /usr/openv/netbackup/bin/bpsched bpsched
-ru oracle -rg dba -ct 4 -ppid 29912 -client UE23 -class RMAN_Hot_ZPRD
-keyword zprd -cv -pl
/usr/openv/netbackup/logs/user_ops/dbext/logs/2272.0.1091767567 -UB
-clnt_lc_messages C -clnt_lc_time C -clnt_lc_ctype C -clnt_lc_collate C
-clnt_lc_numeric C
22:46:08.576 [29912] <2> wait_for_child: start
22:56:47.224 [29912] <2> wait_for_child: child exit_status = 54
signal_status = 0
22:56:47.224 [29912] <2> bkarfiles: User backup failed (client = UE23 user =
oracle group = dba): timed out connecting to client
22:56:47.224 [29912] <2> mail_msg: entered; status = 54
22:56:47.225 [29912] <2> mail_msg: Attempting to send mail to oracle on UE23
22:56:47.225 [29912] <2> mail_msg: Getservbyname returned port 13782
22:56:47.226 [29912] <2> getsockconnected: Connect to UE23 on port 989
22:56:47.244 [29912] <2> logconnections: BPCD CONNECT FROM 10.128.31.126.989
TO 10.128.31.107.13782
22:56:48.257 [29912] <2> mail_msg: CMD_SOCK from bpcr = 5
22:56:48.257 [29912] <2> mail_msg: STAT_SOCK from bpcr = 8
22:56:48.482 [29912] <2> bpcr_get_version_rqst: bpcd version: 04500000
22:56:48.588 [29912] <2> bpcr_get_version_rqst: bpcd version: 04500000
22:56:48.807 [29912] <2> bpcr_get_version_rqst: bpcd version: 04500000
22:56:49.007 [29912] <2> mail_msg: BACKUP EXIT STATUS = 54
22:56:49.008 [29912] <2> getsockconnected: Connect to UE23 on port 774
22:56:49.025 [29912] <2> logconnections: BPCD CONNECT FROM 10.128.31.126.774
TO 10.128.31.107.13782
22:56:49.725 [29912] <2> process_request: EXIT STATUS 54
22:56:49.725 [29912] <2> bprd: /usr/openv/netbackup/bin/userreq_notify
backup UE23 oracle



-----Original Message-----
From: Donaldson, Mark - Broomfield, CO 
Sent: Friday, August 06, 2004 9:32 AM
To: Veritasbu (E-mail)
Subject: Error 54 - client connect timeout


I'm being plagued by these on my RMAN backups:

ERROR: 54
timed out connecting to client
The server could not complete the connection to the client. The accept
system or winsock call timed out after 60 seconds.

I've got an open ticket with Veritas but have any of you defeated this
dragon?

-M

=================================================
  (2B|^2B) == ?
=================================================
  Mark Donaldson - SA - Corporate Express
=================================================