NETGEAR is aware of a growing number of phone and online scams. To learn how to stay safe click here.
Forum Discussion
Solved
Creating new volumes is the resync operation detrimental
I wish to test different RAID levels. First to create a RAID 1 volume using two 4TB disks. On the start of the second month to backup, the data accumulated delete the volume and create a new RAID 6 v...
readynasuser01 wrote:
Could you explain how you worked out 48TB per year so I can apply it to other disks sizes? In my case is it 4TB if done every month for a year is 12 resyncs. So 4Tb x 12 months - 48Tb?
That's it.
readynasuser01 wrote:
How did you get 48Tb - 64Tb per year per drive. If done every 3 months that's 4 scrubbing operations carried out a year. So for a 6Tb disk x 4 (resyncs/scrubbing) = 24Tb. For a 8Tb disk x 4 (resyncs/scrubbing) = 32Tb. Did you double the amount to account for every day read and write operations by network users?
I doubled it because I am also running the disk test every quarter (which also reads every sector). Balance and Defrag also would have some impact, but on my systems they complete quickly, so they aren't doing a lot of disk I/O.
readynasuser01 wrote:
It is backward but another NAS will not be enrolled anytime soon. Believing by testing different RAID levels and getting user feedback can decide what RAID level works best and stick with it for the long run.
I'd try RAID-10 while you are at it. It should be at least as fast as RAID-1, and would give you protection against some 2-drive failure combinations. So it's a middle ground between RAID-1 and RAID-6.
Dual-RAID1 is another option if you need the space, though you would need to balance the storage between the two volumes manually. Protection there is identical to RAID-10. One advantage of staying with RAID-1 is that it has the simplest data recovery (if you ever need it).
It's a ReadyNAS 314. I'm still unclear when a resync operation is done on the same disks frequently is it bad for the hard disks? If the question was answered by saying it is a block-level function, I didn't understand. If the resync operation cause disks to read and write a lot then I know this wear and tear would shorten the life span. Is this what the resync is doing?
readynasuser01 wrote:
If the resync operation cause disks to read and write a lot then I know this wear and tear would shorten the life span. Is this what the resync is doing?
Every sector of the disks is either read or written by the resync. So yes, there is a lot of disk i/o going on.
How much this affects the life span of the drive is an interesting question, and I haven't seen a clear answer to it. Seagate is beginning to put expected disk workloads in their drive manuals. They say that their IronWolf drives are designed to handle 180 TB per year. Their IronWolf Pro drives are designed to handle 300 TB per year. I suspect WDC Reds and Red Pros are designed around similar workloads.
Resyncing every month would require 48 TB per year, which seems ok to me. Resyncing every week on the other hand is 208 TB/year, and I would be concerned about shortening drive life in that scenario.
I scrub and run disk tests every 3 months. I want early warning when the disks begin to fail, so I can replace them quickly. That's the best way I know of to minimize the risk of a double-drive failure. Since I am using 6-8 TB drives, those maintenance functions create a workload of 48-64 TB per year per drive.
readynasuser01 wrote:
If the question was answered by saying it is a block-level function, I didn't understand.
"block level" means that the processing works on the raw disk blocks and doesn't depend on the file system. So the resync processing is the same when the volume is totally empty as it is when the volume is full. It's a fairly common term with storage systems - for example iSCSI is also block level, providing a virtual disk for the iSCSI client. The client needs to format that disk, etc - just as if the disk were connected to the client.
readynasuser01 wrote:
First to create a RAID 1 volume using two 4TB disks. On the start of the second month to backup, the data accumulated delete the volume and create a new RAID 6 volume using 4 4TB disks. Then copy the backup data back onto the NAS.
If I understand this correctly, your test involves two resyncs over the life of the disks. If you like RAID-1 better then you'd have a third resync. 12 TB total i/o per drive worst case, which should have no impact on the drive lifetime. If you added RAID-10 to your test, you'd still only have 16 TB per drive, which also should have no impact.
You might be going about this backwards though. I'd begin by sorting out what data protection / availability I needed, and then get a NAS that provides that with good performance.
Thanks a lot, StephenB that the information given helped to make a decision. The NAS has WDC Reds drives at current. Could you explain how you worked out 48TB per year so I can apply it to other disks sizes? In my case is it 4TB if done every month for a year is 12 resyncs. So 4Tb x 12 months - 48Tb?
How did you get 48Tb - 64Tb per year per drive. If done every 3 months that's 4 scrubbing operations carried out a year. So for a 6Tb disk x 4 (resyncs/scrubbing) = 24Tb. For a 8Tb disk x 4 (resyncs/scrubbing) = 32Tb. Did you double the amount to account for every day read and write operations by network users?
It is backward but another NAS will not be enrolled anytime soon. Believing by testing different RAID levels and getting user feedback can decide what RAID level works best and stick with it for the long run.
Thanks.
readynasuser01 wrote:
Could you explain how you worked out 48TB per year so I can apply it to other disks sizes? In my case is it 4TB if done every month for a year is 12 resyncs. So 4Tb x 12 months - 48Tb?
That's it.
readynasuser01 wrote:
How did you get 48Tb - 64Tb per year per drive. If done every 3 months that's 4 scrubbing operations carried out a year. So for a 6Tb disk x 4 (resyncs/scrubbing) = 24Tb. For a 8Tb disk x 4 (resyncs/scrubbing) = 32Tb. Did you double the amount to account for every day read and write operations by network users?
I doubled it because I am also running the disk test every quarter (which also reads every sector). Balance and Defrag also would have some impact, but on my systems they complete quickly, so they aren't doing a lot of disk I/O.
readynasuser01 wrote:
It is backward but another NAS will not be enrolled anytime soon. Believing by testing different RAID levels and getting user feedback can decide what RAID level works best and stick with it for the long run.
I'd try RAID-10 while you are at it. It should be at least as fast as RAID-1, and would give you protection against some 2-drive failure combinations. So it's a middle ground between RAID-1 and RAID-6.
Dual-RAID1 is another option if you need the space, though you would need to balance the storage between the two volumes manually. Protection there is identical to RAID-10. One advantage of staying with RAID-1 is that it has the simplest data recovery (if you ever need it).
NETGEAR Academy
Boost your skills with the Netgear Academy - Get trained, certified and stay ahead with the latest Netgear technology!
Join Us!
