Check your lan settings.  All ports should be full-duplex auto-neg OFF.
Auto-neg or half-duplex will KILL performance.


|---------+--------------------------------------->
|         |           Elwood Blues                |
|         |           <fernando AT secret DOT org>       |
|         |           Sent by:                    |
|         |           veritas-bu-admin AT mailman DOT eng|
|         |           .auburn.edu                 |
|         |                                       |
|         |                                       |
|         |           04/04/02 02:40 PM           |
|         |                                       |
|---------+--------------------------------------->
  
>---------------------------------------------------------------------------------------------------------|
  |                                                                             
                            |
  |       To:       <Veritas-bu AT mailman.eng.auburn DOT edu>                  
                                   |
  |       cc:                                                                   
                            |
  |       Subject:  [Veritas-bu] Slow backups, but only on one client           
                            |
  |                                                                             
                            |
  |                                                                             
                            |
  
>---------------------------------------------------------------------------------------------------------|



A customer is having a problem with one of their servers backups. Running
NB Datacenter server on Solaris 7, with several Solaris 7 & 8 clients. All
the clients backup fine, except for one (Solaris 8, UltraSPARC-IIi 440MHz,
768M RAM). The only thing different about this client is that it is a mail
server. If the backup is allowed to run, it will sometimes take up to 4
days to complete, blocking out other NetBackup clients during that time.
Initially we were assuming that the backups were taking a long time due to
constantly updating mail spools and mqueue directories. We altered what
files would be backed up, limiting it to config files and binaries, and
backups would still take forever. Looking at the log files during a backup
showed very little information, the backup would sit for 30 or more
minutes at a time backing up something small like an /etc/rc*.d file. The
server has plenty of available cpu, memory and swap space. These servers
all hang off the same VLAN, and the server that is having difficulty has
plently of available bandwidth. The other NetBackup clients are backing
up significantly more data in much less time (15-45 minutes).

Can anyone shed any light on the situation, any suggestions for what kind
of additional logging I should enable, etc?

TIA

    -F

_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu