Legato Networker timing paratmer

Hello all,
One simple question,in veritas,there is a parameter
MEDIA_UNMOUNT_DELAY,does Legato networker have a function like
parameter,what is it?
Email answer is preferred.

Thanks
Victor

Well, Legato uses the following sleeps/timeouts:

1) Configurable in the jukebox:
Eject sleep: Define the number of seconds for a jukebox to sleep after
an eject operation.
Unload sleep: Define the number of seconds for a jukebox to sleep
before attempting to unload a given cardridge.
Idle device timeout: The number of minutes before an idle device
is unmounted (only in shared device
environments).

2) Configurabe in the group:
Inactivity timeout: The number of minutes of inactivity before the
group of
saves concludes that a client mashine is
(presumably)
permanently hung.

Rene

"Victor Zhang" schrieb im Newsbeitrag
om...
Hello all,
One simple question,in veritas,there is a parameter
MEDIA_UNMOUNT_DELAY,does Legato networker have a function like
parameter,what is it?
Email answer is preferred.

Thanks
Victor

Rene,
Thanks for your reply.
One more question,why is there a need for these delay/sleep setting?
Why shouldn't hardware notify software the ending of cartridge
ejection,unloading,and other activity?

Regards
Victor

One more question,why is there a need for these delay/sleep setting?
Why shouldn't hardware notify software the ending of cartridge
ejection,unloading,and other activity?

In fact, hardware does this notification you're talking about via normal

scsi-commands (command completion message and status good).

As far as I know, software developers decided to add some additional

delays only for safety reasons. Guess they've made thousands of test

with different hardware and got sporadically errors if they only count

on hardware notifications.

Additional, hardware notifications are sometimes improper in

environments with some forms of caching, and with multiple

command buffering.

Decreasing those delays should be done very careful, my recommendation

is to ask the vendor of your hardware about it.



Rene

Rene,
Thank you very much.

Regards
Victor
"Rene Köhnen-Wiesemes" wrote in message ...
One more question,why is there a need for these delay/sleep setting?
Why shouldn't hardware notify software the ending of cartridge
ejection,unloading,and other activity?

In fact, hardware does this notification you're talking about via normal

scsi-commands (command completion message and status good).

As far as I know, software developers decided to add some additional

delays only for safety reasons. Guess they've made thousands of test

with different hardware and got sporadically errors if they only count

on hardware notifications.

Additional, hardware notifications are sometimes improper in

environments with some forms of caching, and with multiple

command buffering.

Decreasing those delays should be done very careful, my recommendation

is to ask the vendor of your hardware about it.



Rene