Hi,
Following a data loss scare, I am in the process of planning the conversion of my current HW Raid* to ZFS, and have been looking at FreeNAS as a solution.
I wanted to validate my thought process for backup before I go ahead with building/testing anything.
Though I have quite a bit of data ~15TB, only about 3TB of it can be classified as "invaluable", the rest consists of the CD/DVD/Blu-ray rips which can easily (though time consuming) be re-ripped. Though I have briefly looked into building a second server, this is cost prohibitive right now, so I will have to make do with what I have and backup locally, rotating the disks to store off-site.
My important data currently sits on a RAID6 array of 6x2TB HGST NAS disks which I plan to convert to a RAID-Z2 array (I have sufficient capacity in other arrays to copy the data off and then back again after the conversion). I also have 6x1.5TB WD Green disks which I plan on using for a local backup of this main array using zfs send/receive for an incremental backup.
I had originally planned on configuring the backup array as a 4x1.5TB RAID10 mirror and then scripting mounting/detaching the mirrored set and rotating this on a regular basis. Would I be right in my conclusion that (aside from being quite difficult to script) that the mirror wouldn't just pick-up from where it last left off in the write process but would blank everything on the disk and then sync with the static set?
Would I therefore be better off creating two 3 disk Raid-z1 pools, and scripting a mount/detach and backup to these?
Finally, I wanted to ask a CrashPlan question. I've seen on the forum that Crashplan can be installed as a plug-in and configured to back-up to CrashPlan central, however I also use my NAS as a backup destination for the other computers in my house. CrashPlan has it's own archive format, which I believe is compressed and handled somewhat like a database file. Are there any special considerations for the dataset where I configure these backup stores to go, for example turning off compression or the cow features?
Thanks in advance,
~Dan
Build details:
Intel Xeon E3-1230 V2 | Supermicro X9SCM-F | 32GB Crucial DDR3 1333 ECC | LSI 9240-8i (Flashed to IT mode) | HP 32 Port SAS Expander
6x2TB HSGT NAS (data) | 6x1.5TB WD Green (backup) | 6x3TB HSGT (media 1) | 3x2TB Seagate (media 2)
*HW Raid controller is an Areca 1882LP - this will be going in the new world.
Following a data loss scare, I am in the process of planning the conversion of my current HW Raid* to ZFS, and have been looking at FreeNAS as a solution.
I wanted to validate my thought process for backup before I go ahead with building/testing anything.
Though I have quite a bit of data ~15TB, only about 3TB of it can be classified as "invaluable", the rest consists of the CD/DVD/Blu-ray rips which can easily (though time consuming) be re-ripped. Though I have briefly looked into building a second server, this is cost prohibitive right now, so I will have to make do with what I have and backup locally, rotating the disks to store off-site.
My important data currently sits on a RAID6 array of 6x2TB HGST NAS disks which I plan to convert to a RAID-Z2 array (I have sufficient capacity in other arrays to copy the data off and then back again after the conversion). I also have 6x1.5TB WD Green disks which I plan on using for a local backup of this main array using zfs send/receive for an incremental backup.
I had originally planned on configuring the backup array as a 4x1.5TB RAID10 mirror and then scripting mounting/detaching the mirrored set and rotating this on a regular basis. Would I be right in my conclusion that (aside from being quite difficult to script) that the mirror wouldn't just pick-up from where it last left off in the write process but would blank everything on the disk and then sync with the static set?
Would I therefore be better off creating two 3 disk Raid-z1 pools, and scripting a mount/detach and backup to these?
Finally, I wanted to ask a CrashPlan question. I've seen on the forum that Crashplan can be installed as a plug-in and configured to back-up to CrashPlan central, however I also use my NAS as a backup destination for the other computers in my house. CrashPlan has it's own archive format, which I believe is compressed and handled somewhat like a database file. Are there any special considerations for the dataset where I configure these backup stores to go, for example turning off compression or the cow features?
Thanks in advance,
~Dan
Build details:
Intel Xeon E3-1230 V2 | Supermicro X9SCM-F | 32GB Crucial DDR3 1333 ECC | LSI 9240-8i (Flashed to IT mode) | HP 32 Port SAS Expander
6x2TB HSGT NAS (data) | 6x1.5TB WD Green (backup) | 6x3TB HSGT (media 1) | 3x2TB Seagate (media 2)
*HW Raid controller is an Areca 1882LP - this will be going in the new world.
