Hi,
Twice a month I'm backing up 9 TB of data. I have 2 text files which include about 4000 files each.
Those files are backed up just with a save command on the (dedicated) Storage Node:
save -s backup_server -N R1 -w "2 month" -y "2 month" -b pool_name -I R1
save -s backup_server -N R2 -w "2 month" -y "2 month" -b pool_name -I R2
On the Storage Node this backup takes +-2 days to complete.
However the restore is pretty simple:
1. find the 2 SSID's:
mminfo -q "name=R1,name=R2" -r client,savetime,name,ssid
2. Run the recover command
recover -s backup_server -S SSID_OF_R1
recover -s backup_server -S SSID_OF_R2
Now with the above procedure, only one session is used during the backup, and that's what I'm trying to change.
I've created a pool, a client (Which is actually a Dedicated Storage Node), 2 groups (for R1 and R2).
I've written a script which fills up the save set list in the client configuration. This is because during each backup (twice a month) the lists R1 and R2 always changes.
I believe this can speed up the backup a lot because the Storage Node has access to several tape drives and can handle lots of simultanious sessions.
On the other hand, if I have to restore everything later, I'll have to deal with thousands of save sets. Using the recover command with lots of "- S" options will just restore the files one by one (one restore session only) and so it could takes days to complete ... in best case.
Can someone help me out with my choice of backup, or advise another way to backup/restore with the highest performance for backup and resotre taken into consideration?
Backup Server and Client (Storage Node) Both NW 7.4.4 running on RHEL5.2
thanks
GS
Twice a month I'm backing up 9 TB of data. I have 2 text files which include about 4000 files each.
Those files are backed up just with a save command on the (dedicated) Storage Node:
save -s backup_server -N R1 -w "2 month" -y "2 month" -b pool_name -I R1
save -s backup_server -N R2 -w "2 month" -y "2 month" -b pool_name -I R2
On the Storage Node this backup takes +-2 days to complete.
However the restore is pretty simple:
1. find the 2 SSID's:
mminfo -q "name=R1,name=R2" -r client,savetime,name,ssid
2. Run the recover command
recover -s backup_server -S SSID_OF_R1
recover -s backup_server -S SSID_OF_R2
Now with the above procedure, only one session is used during the backup, and that's what I'm trying to change.
I've created a pool, a client (Which is actually a Dedicated Storage Node), 2 groups (for R1 and R2).
I've written a script which fills up the save set list in the client configuration. This is because during each backup (twice a month) the lists R1 and R2 always changes.
I believe this can speed up the backup a lot because the Storage Node has access to several tape drives and can handle lots of simultanious sessions.
On the other hand, if I have to restore everything later, I'll have to deal with thousands of save sets. Using the recover command with lots of "- S" options will just restore the files one by one (one restore session only) and so it could takes days to complete ... in best case.
Can someone help me out with my choice of backup, or advise another way to backup/restore with the highest performance for backup and resotre taken into consideration?
Backup Server and Client (Storage Node) Both NW 7.4.4 running on RHEL5.2
thanks
GS
