09-02-2015 07:57 AM
I have had issues in the past with SQL database restores, I run my backups with multiple child jobs, but I have found when I restore that I have to run as a single thread. This means that a backup of 4 child jobs, each taking an hour, backs up in one hour, but restores in four.
To allow for faster restores, we have set up SQL backup directly to disk,and restore from there. We make backups of these flat files for DR purposes.
I am currently on 7.6.0.4 - does anyone know if this issue remains in the newer versions of NetBackup? 7.6.1 or 7.7?
09-02-2015 08:15 AM
Interesting, curious if the threads/child jobs mentioned relates to stripes or databases in the SQL instance.
09-02-2015 01:20 PM
I believe this is stripes. There are multiple databases being backed up, usually only one being restored.
Have you been able to restore using multiple stripes?
09-02-2015 02:01 PM
Use script to perform the restore and add STRIPES 4 to it and try.
09-02-2015 11:41 PM
Have not tried restoring with stripes, as the infrastructure already gave around us around 70 MB/sec at restore which was fast enough for us. We found that the limit in our system most often was the disk we was writing to.