Help with a multitude of Brightstor issues

[vancleb]

I have a Brightstor for NetWare install r11.1 SP1 which is on a new NetWare 6.5 server.
I want to move data from an older NetWare 5.1 server to this new server and purchased a NetWare Agent to do so.
I attempted to "Copy" from the original server to the new server but can't see the original server in the source list even witht he agent loaded and I have IP communication between the two server without problem, they are on the same 100MB switch. Don't know how to make the original visable to the manager to start the copy.
Because of this I opted to backup and restore. The backups worked without issue. In the restore, I am restoring by session to selectively restore various directories to certain places on the new server but can't with the session highlighted, I don't seen any of the folders or files, I have to restore all then move.
Lastly.... The restore of these entire volumes are Painfully slow even thought the tape device is directly attached to the server I am restoring to. I backed up witht he same configuraton and the backup over the network was 10 times faster than the restore is going ...
Any help greatly appriciated

 

[vschumpy]

1 Copy jobs don't use the agent.

2 The restore speed is almost certainly a NetWare IP stack limitation.

 

[ShineOn]

The appropriate method for moving data from one NetWare server to another is to use the Server Consolidation utility, available as a free download from Novell. It will maintain trustee rights for you, automatically modifying them to the new server/volume location.

Backup software is not built for migration, either within platforms or cross-platform.

Go to

http://download.novell.com

and look for "Server Consolidation Migration Toolkit", filename SCMT.EXE, which has both the latest Migration Wizard and the Server Consolidation Utility.

 

[ShineOn]

Also, restore speed isn't a NetWare IP stack limitation - it's a Windows IP stack limitation (I say this because the Windows IP stack is what the ARCserve communications protocols are based on - not the "real world" of IP.)

If you turn off nagle algorithm and delayed ack just for during the restore, the restore will be fast.

If you don't turn delayed ack and nagle back on afterward, it'll cause problems with normal communications.

I like to also turn on Minshall's algorithm for a "middle-ground" approach, rather than turning off delayed ack and Nagle.

 

[vschumpy]

Not true. A majority of users in this situation will have zero windows to windows restore speed issues but always have slow remote netware restore issues (and this has nothing to do with a different buffer size being used or passed from windows). In fact the same can behaviour can usually be found if you are using a netware host server install and nwagent on a remote machine.

In the past this could be gotten around by to be the old nagle algorithm or using TCP delayed ack, or other TCP set commands, or using the Null or export TCPIP NLM version. IIRC this has been done automatically by the NWAGENT on the fly when jobs run since v9.0 (or maybe 11.x - can't remember off the top of my head), and set back again afterwards. I've never found toggling minshalls to have any significant effect on restore speed.