I understand that off-the-shelf NASs can sync to services like Backblaze, but the potential volume concerns me here. Off-site backups: No real plan has taken shape yet. QNAP seems to have a ZFS enterprise product, but it's a rackmount, and we don't have racks. I don't exactly trust btrfs, even with Synology's official support. Ideally I'd use ZFS, but that seems in conflict with going off-the-shelf. Because of that, I'm inclined to go with an off-the-shelf.īit rot: I am modestly concerned about whether this will be an issue at this scale. Moreover, the expected service life will be long, and the system needs to be reasonably maintainable in the long term. Off-the-shelf: No one, including myself, has previous experience building NAS's. I prefer HGST based off their track record, both in personal use and from Backblaze's data.ĭIY vs. 8 TB NAS disks seem sensible, and all three major vendors (WD, HGST, and Seagate) seem to be in the ~$270/drive range. I think this is overly pessimistic if we do use a RAIDZ2-like configuration.ĭisk/Pool arrangement: To be determined, but potentially 2 x RAIDZ2 of 6 x 8TB, for 64GB usable. Usable space: The estimated actual working space needed is ~50TB, assuming a factor of 2 for parity. My biggest worries are with respect to bit rot / data integrity and the disk/pool arrangement. I'd like to see if I'm aware of the main considerations here, and if this overall makes a reasonable amount of sense. The budget is accordingly in the ~$6k range. a Synology DS3617xs) filled with 8 TB NAS disks. While I'm reasonably aware of trends in personal NAS's, the size of the array being considered - 100 TB - is dramatically different from what I had worked out for myself.Īs it is, the working idea is a very large off-the-shelf NAS (e.g.
My organization is considering a NAS to store a mixture of data, ranging from short-term working backups to large, long-term archival data for preservation.
0 kommentar(er)
