This post is more than 5 years old
3 Posts
0
486
Can I change my dataset to create an aggregate backup?
I am new to Avamar and I have been recently charged with taking on a large enterprise implementation. I have a server that has has 8 1TB drives. 2 of them are full and 2 other are starting to fill up. There are around 16 million files. Avamar cannot complete a full backup of the system in our backup window and it fails with a time out error. Since a large number of the files never change I assume that after I have a full backup that Avamar will only backup the new changes and it will be able to complete within the window.
So my question is this...can I set the dataset to backup drives 1-4 today then set it to backup 5-8 tomorrow and then set it to backup everything after that? My theory is that after 2 days I would have a full backup of all 8 drives and then I will only be backing up the changes after that.
I may be over thinking this or completely off but I am trying to avoid running a full backup through the blackout period because we have TONS of other stuff backing up and I need the maintenance to run its course. I am also trying to avoid creating multiple groups and datasets for one client.
Anyway I would appreciate any help you could provide me on this so that I can get this client a full backup and then continue to backup the changes from here on out.
ionthegeek
2K Posts
0
June 17th, 2013 14:00
Normally the Avamar client only permits one workorder at a time to run on a client but there are ways to work around this if necessary. This is really the "nuclear" option for performance troubleshooting, though and it involves a certain amount of management overhead so I would recommend getting in touch with support to evaluation the situation first. They can take a look at the system, run some performance tests, give you appropriate recommendations, and help you with the configuration if you run into problems.
Seekor
3 Posts
0
June 17th, 2013 14:00
Well that's good and bad news, our window is 13 hours. Would it be best to create a second group and dataset for this client and then backup half the drives in each group? I don't mind having 2 groups and datasets I think I am just a little hesitant to make changes due to my lack of knowledge using Avamar.
ionthegeek
2K Posts
0
June 17th, 2013 14:00
Absolutely! This is one of the strategies I regularly present to the support team and customers if they have large datasets that need to be backed up -- start with a relatively small dataset and build your way up.
Keep in mind, however, that the Avamar client can typically "stat" (i.e. read the metadata for) approximately 1 million files per hour and the client needs to "stat" each file daily in case it has changed since the last backup. If your backup window is shorter than 16 hours you may need to work with the support team to do some performance optimization.
ionthegeek
2K Posts
0
June 17th, 2013 14:00
Also, FYI, NDMP backups do not have this limitation on the number of unchanged files since the client asks the NAS device for a list of the files that have changed since the last backup (so it's not necessary to "stat" the unchanged files).
Seekor
3 Posts
1
June 17th, 2013 14:00
Thanks for your help, it will definitely get me moving in the right direction.
ionthegeek
2K Posts
0
June 18th, 2013 07:00
My pleasure!