I was trying to restore some data from a Solaris 9 dedicated storage node to a Solaris 10 client. Since the Solaris 10 is a client then that means that the recover will go through the network. Everything was fine until I got into one subfolder for a user which failed giving me:
recover command started on Wed Mar 11 09:28:33 2009
Initiating remote recover to directed_clienter from server backuper, this may take a while…
39078:recover: SYSTEM error: Connection timed out
16200:recover: Unable to send directed recover file /path/to/folder/filename
39078:recover: SYSTEM error: Invalid argument
After looking at the data, it appeared to be a lot of files in this subdirectory (66 pages on the nwrecover windows plus a few directories).
I contacted support and told me try a local recovery because it looks like a network problem. As simple as that, it worked just fine! The guy was telling me that the connection died out waiting for a response from the backup server because there were so many files indexed. He said that restoring a saveset might have worked fine because it doesn’t look at the indexes.
So in short, backing up to the backup server locally worked just fine and I’ll probably just nfs mount the directed client server locally so I can avoid this problem.
P.S: using networker 7.4.4