I am somewhat new to FreeNAS, getting my feet wet, trying things out, and seeing if this system is right for our environment.
Our current storage server is running FreeBSD 9 with ZFS on geli-encrypted drives for storage. We run some scripts on the box to periodically go fetch new files from several servers, and also daily to back up all our servers. I am fond of pulling data this way versus the servers pushing data into the box, due to the lower exposure this gives the backup server from the outside world.
I am setting up something similar to this under FreeNAS, in the hope that we can replace the FreeBSD box with FreeNAS. As I understand it, I need to create at least one jail to house and run our backup scripts.
Where I am starting to question if moving to FreeNAS is appropriate for us is in how jails access storage outside the jailed filesystem. Since we have at least a dozen different datasets for the various backups, these will need to be added individually to the jail. I assume this is technically possible, but at this point I wonder if it is wise to do this. Is there a performance penalty for accessing the dataset through the jail storage system this way? Are there risks to having so many datasets mounted inside the jail?
And more importantly, am I going about this the wrong way? Can I do the same thing without jails?
Thanks for your input and advice.
Our current storage server is running FreeBSD 9 with ZFS on geli-encrypted drives for storage. We run some scripts on the box to periodically go fetch new files from several servers, and also daily to back up all our servers. I am fond of pulling data this way versus the servers pushing data into the box, due to the lower exposure this gives the backup server from the outside world.
I am setting up something similar to this under FreeNAS, in the hope that we can replace the FreeBSD box with FreeNAS. As I understand it, I need to create at least one jail to house and run our backup scripts.
Where I am starting to question if moving to FreeNAS is appropriate for us is in how jails access storage outside the jailed filesystem. Since we have at least a dozen different datasets for the various backups, these will need to be added individually to the jail. I assume this is technically possible, but at this point I wonder if it is wise to do this. Is there a performance penalty for accessing the dataset through the jail storage system this way? Are there risks to having so many datasets mounted inside the jail?
And more importantly, am I going about this the wrong way? Can I do the same thing without jails?
Thanks for your input and advice.
