Hi Brian,
Recently I asked the same question to Legato.
Legato has performed a test for this, and the summary is presented below:
Summary:
1.
Solaris system with Networker 6.1 Build 186 as Backup Server
Configured with 256 local file devices
Target session is changed from 4 to 1 to allow for one data stream to one
device.
Test backup is consisted of 256 bigasm save sets, 500MB each.
Finding: The system was able to write 256 devices simultaneously.
2.
Same system, Networker was then upgraded to 6.1.1 Build 238.
Bug LGTpa33271/LGTpa36836 was opened because nsrd was in a loop and could not
service any requests when NetWorker was started with 256 mounted volumes.
LGTpa33273 was filed because NetWorker 6.1.1 could not complete a backup
with 300
save sets when 256 devices were active. This test ran successfully on 6.1
3.
The last test was performed
on a Solaris system configured as a NetWorker server with no local devices and
a separate Solaris system configured as a NetWorker storage node with 256
file type devices.
The results were the same for 6.1 and 6.1.1.
Nsrd died when it created the 183rd remote device.
Since nsrd died, no further testing was performed and
LGTpa33288/LGTpa36780 and LGTpa36781 were created for this bug.
Engineering has resolved and checked in the fixes for
LGTpa33271/LGTpa36836, LGTpa36781 and
LGTpa33288/LGTpa36780 into the dev tree and the fixes have been ported to
the 6.1.2 release and 7.0
branch.
Once LGTpa33288/LGTpa36780 were fixed, a new test was executed.
255 file type devices were created on the storage node and 1 file type
device was created on the
NW server for the bootstrap.
Then a backup containing 256 bigasm save sets of 500MB each was executed
and ran successfully.
This test had all 255 remote devices active simultaneously.
hope it answer your queries.
cheers - han
At 11/15/2002 01:36 PM, Brian Plume wrote:
Good afternoon, everyone.
I seem to recall reading a white paper on a software limitation of 45
nsrmmd processes for a windows based networker server. Is this still
true? If so, how are you people backing up SAN attached servers? Given
the example:
7 SAN storage nodes all seeing a 6 drive tape library, that would be a
total of 42 nsrmmd processes. That's CLOSE to the 45 limit I recall
seeing, and is a VERY SMALL amount of servers to back up through a
SAN. Am I missing something here, or is the newest version of Networker
supporting more processes?
Thanks in advance. This is my first posting to this group. Enjoy your day.
Brian Plume
Enterprise Storage Architect
Nth Generation Computing, Inc.
Phone: 858-451-2383 Fax: 858-673-8431
"Service and Technology to the Nth Degree"
http://www.nth.com <http://www.nth.com/>
--
Note: To sign off this list, send a "signoff" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
--
Note: To sign off this list, send a "signoff" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
cheers - han
|
