Reply
Topic Options
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page
how does the EDA500 work? and safety from disk failures
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
2013-05-16
05:33 AM
2013-05-16
05:33 AM
how does the EDA500 work? and safety from disk failures
Hello,
I'm wondering how the EDA500 works. Does it simply provide more bays as if they were incorporated in the base unit chassis, or can you have an additional protected set of disks? (for example, with a 316 and 500 would you have all 11 disks as one xraid set, or can you have a 4disk+2parity set on the 316 and put a 4disk+1parity set on the EDA and have them both appear as one volume?)
My concern is protecting against drive failures. My impression is that with a 1-parity setup, the safest configuration is a 3 spindle set (2+1parity), and extra non-parity disks just simply increases the chance that two drives could fail before parity could be reestablished.
Or is maintaining a high-parity amount not a concern? (Are most problems simply misread bits, handled via btrfs' checksumming?)
Thank you.
I'm wondering how the EDA500 works. Does it simply provide more bays as if they were incorporated in the base unit chassis, or can you have an additional protected set of disks? (for example, with a 316 and 500 would you have all 11 disks as one xraid set, or can you have a 4disk+2parity set on the 316 and put a 4disk+1parity set on the EDA and have them both appear as one volume?)
My concern is protecting against drive failures. My impression is that with a 1-parity setup, the safest configuration is a 3 spindle set (2+1parity), and extra non-parity disks just simply increases the chance that two drives could fail before parity could be reestablished.
Or is maintaining a high-parity amount not a concern? (Are most problems simply misread bits, handled via btrfs' checksumming?)
Thank you.
Message 1 of 3
Labels:
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
2013-05-16
05:45 AM
2013-05-16
05:45 AM
Re: how does the EDA500 work? and safety from disk failures
You would have a separate volume on the EDA500 e.g. so you could backup the data in the main unit via backup job(s)
Message 2 of 3
- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
2013-05-16
10:08 AM
2013-05-16
10:08 AM
Re: how does the EDA500 work? and safety from disk failures
checksumming doesn't recover misread bits, all it does is allow the problem to be detected. So you'd still depend on the underlying RAID for repair. The checksum in principle could let you tell which block was actually bad (since you can regenerate each block from the others, and see which one results in the correct checksum).
In all raid arrays, the risk of failure rises as the number of disks in the data volume goes up. RAID-1 pairs are better than RAID-5 for protection, but they also reduce the potential volume size. So there is a tradeoff.
I'm not sure if I would use an EDA500 as a backup for the main NAS or not - I think I'd rather have the backup on a completely different device, and use the EDA500 for expansion. Though I wouldn't want to put the EDA500 disks into the same array as the main NAS. Separate volumes (as mdgm suggests) seems safer.
In all raid arrays, the risk of failure rises as the number of disks in the data volume goes up. RAID-1 pairs are better than RAID-5 for protection, but they also reduce the potential volume size. So there is a tradeoff.
I'm not sure if I would use an EDA500 as a backup for the main NAS or not - I think I'd rather have the backup on a completely different device, and use the EDA500 for expansion. Though I wouldn't want to put the EDA500 disks into the same array as the main NAS. Separate volumes (as mdgm suggests) seems safer.
Message 3 of 3