Poor performance on directet recovery

Try restoring a small save set. It seems the drive is not reading the tape.

What is indicated on the Library panel?

I won't agree that TSM is better than Networker. One odd experience should not be a cause to form an opinion. NW 7.3.2 is also not bad.

Try and restore the entire stuff again and what is the processing power and memory on the backup server available at the time you are recovering the large data?

I stopped the client session restore but the process remains on Legato server so I had to kill it.

root 17447 1 0 Oct19 ? 00:21:07 /usr/sbin/nsrexecd
root 17456 1 0 Oct19 ? 00:21:56 /usr/sbin/nsrd
root 17473 17456 0 Oct19 ? 00:03:56 /usr/sbin/nsrmmdbd
root 17476 17456 0 Oct19 ? 00:00:10 /usr/sbin/nsrindexd
root 17482 17456 0 Oct19 ? 00:01:20 /usr/sbin/nsrmmgd
root 17490 17456 0 Oct19 ? 00:15:41 /usr/sbin/nsrjobd
root 17606 17447 0 Oct19 ? 00:00:43 /usr/sbin/nsrlcpd -s rpblegato1srv.rompetrol.org -N 1 -n 1
root 17895 17456 3 Oct19 ? 1-07:15:56 /usr/sbin/nsrmmd -n 1
root 17900 17456 1 Oct19 ? 14:25:40 /usr/sbin/nsrmmd -n 2
root 17905 17456 0 Oct19 ? 07:27:28 /usr/sbin/nsrmmd -n 3
root 17915 17456 0 Oct19 ? 01:48:04 /usr/sbin/nsrmmd -n 5
root 15028 17456 0 Nov06 ? 00:19:59 /usr/sbin/nsrmmd -n 6
root 6294 17456 74 15:20 ? 01:31:39 /usr/sbin/ansrd aix4.rompetrol.org:root recovering
root 6312 17456 0 15:20 ? 00:00:00 /usr/sbin/nsrmmd -n 7
root 6329 17456 0 15:21 ? 00:00:00 /usr/sbin/nsrmmd -n 8
root 6798 5271 0 17:23 pts/0 00:00:00 grep nsr
[root@rpblegato1srv logs]# kill 6294
[root@rpblegato1srv logs]# ps -ef | grep nsr
root 17447 1 0 Oct19 ? 00:21:07 /usr/sbin/nsrexecd
root 17456 1 0 Oct19 ? 00:21:56 /usr/sbin/nsrd
root 17473 17456 0 Oct19 ? 00:03:56 /usr/sbin/nsrmmdbd
root 17476 17456 0 Oct19 ? 00:00:10 /usr/sbin/nsrindexd
root 17482 17456 0 Oct19 ? 00:01:20 /usr/sbin/nsrmmgd
root 17490 17456 0 Oct19 ? 00:15:41 /usr/sbin/nsrjobd
root 17606 17447 0 Oct19 ? 00:00:43 /usr/sbin/nsrlcpd -s rpblegato1srv.rompetrol.org -N 1 -n 1
root 17895 17456 3 Oct19 ? 1-07:15:56 /usr/sbin/nsrmmd -n 1
root 17900 17456 1 Oct19 ? 14:25:40 /usr/sbin/nsrmmd -n 2
root 17905 17456 0 Oct19 ? 07:27:28 /usr/sbin/nsrmmd -n 3
root 17915 17456 0 Oct19 ? 01:48:04 /usr/sbin/nsrmmd -n 5
root 15028 17456 0 Nov06 ? 00:19:59 /usr/sbin/nsrmmd -n 6
root 6312 17456 0 15:20 ? 00:00:00 /usr/sbin/nsrmmd -n 7
root 6329 17456 0 15:21 ? 00:00:00 /usr/sbin/nsrmmd -n 8

After that I've started a single folder from that save set and it seems it's working.
Anyway I must tell you that Networker is not quite a reliable solution for backup. I'm comming from TivoliStorage Manager from IBM and never ever had that numbers of issues

Could be the 7.3.2 version the problem ?

Actually it is, it is just that version 7.3.x so far didn't meet expectations. Tell me, did you go straight to 7.3.x and if not what was your reason to to go for 7.3.x?

Anyway, back to the subject... tested directed restore (HPUX to HPUX) and got 60Mb/s (what was expected). What is your OS and can you elaborate network path and network settings you use?

I'm comming from a TSM environment, Networker is relatively new to me (1 month) ... I must by angree because the recover process did't work as expected but I must tell you with TSM I've never encounter these problems. Could be my fault or something is wrong with my configuration.
I agree with you Networker is a very flexible piece of software but it has faults as any other backup software.

Back to my problem, I've noticed there are 2 save sets for restore, one ~230GB and one 3 GB.
Restoring both save sets with the same session stucks, it loads 3 tapes for the first save set, then issue the message waitng to load another tape for the second save set (I have only 3 drives) and waittttttt ... nothing happened.
During this time on the networker server the ansrd process takes 100% CPU ... but nothing happened. I've waited for 1 hour and nothing.

I killed this process and start the recovery process separately, first the 230GB save set and after that the 3 GB save set. During these recoveries everithing was ok and the recovery speed was as expacted.

Could be a glitch unless you are able to reproduce it. It you are able to reproduce it I'd like to see what happens if you start ssid restore (recover -S) to a box. ansrd thing seems to indicate something could be wrong with restore session itself.