Unsolved
This post is more than 5 years old
1 Message
1
1144
Avamar backups in a RedHat Cluster
We have a large Avamar installation which will be backing up windows systems, VMware infrastructure, bladeserver infrastructure, standalone linux servers and clustered linux servers. Most systems run an Oracle DB.
In a RedHat Linux cluster environment, with relatively "small" operating system file structures, and relatively "LARGE" application & DB file structures(within redHat cluster packages), it "appears" that when a cluster package is moved from one cluster node to another, the Avamar backup is essentially a "new" backup.
My "theory" is that the host's hash-cache content is most meaningful when cluster packages don't move, and that when a cluster package (and all it's filesystems) is moved to a cluster node, that node's hash-cache most likely has little or no contents that matches the chunks of the files and file system that just appeared.
Thus, every time we move a cluster package, it's like a new "first time" backup occurs for the contents.
Please educate me on the specific nitty gritty on how this functions, and if I'm understanding it's operation
Does the 20 byte SHA HASH include any host-specific information
Thanks in advance
Nate Mann