Improve file select performance in NWRECOVER?

[goony]

I just participated in a successful disaster recovery drill (Legato Networker V7.3.2 on Solaris, 9 client systems).

One killer item though was the time it took for the NWRECOVER gui interface (launched from one of the clients) to select all of the files in an incredibly bloated Oracle directory tree (1.16 million files)... from the time the checkbox for the filesystem was clicked until it was finished took over 5 hours!

I'd like to scrap doing this and just do a massive saveset recovery for that directory from the command line, but I'm under pressure to fully document the procedures and "keep them simple" (yea, right).

Is there anything there any way to speed up this file selection via the gui, or suggested alternative (other than the direct saveset recovery that I mentioned)?

Thanks!

 

[605]

Unfortunately, NW takes a very long time to hash the filenames in preparation for the recovery. And i do not know any way to improve this.

The only workaround i know is not to use the file index and run a save set recovery. And it is not difficult.

 

[denisfr]

Have you tried the same thing via the command line 'recover' and 'add' ? Because it is character based (not X), it may be less slow

 

[605]

Don't waste your time - there will be no improvement as far as i remember.

 

[goony]

Thanks for all of the tips...

Command line might save .01 seconds, that's it. While you are waiting for the zillion files to be selected the graphical window doesn't do anything at all, so it's pretty much the same as the command line request.