NETGEAR is aware of a growing number of phone and online scams. To learn how to stay safe click here.
Forum Discussion
Scrub/Defrag/Snapshots with Seagate Archive 8TB - ST8000AS0002
Hi, I am thinking at putting a pare of these disks as XRAID (data redundancy) in my legacy Ultra 2 unit with latest 6.4.1 firmware. I understand the limitations. However, I does own one in my de...
Thanks, I will wait a little as you suggested
I also read about the issues, fixed in 6.4.1 from the feedback
This disk is in fact very tempting for the price /Gb
For archive/backup purpose, in my desktop, it is just an amazing disk with great performances.
I have one Archibe 8Tb disk in my desktop PC since 6 months now. I use it as a backup that is mirrored to the NAS each 6h on a schedule. Performance wise, the NAS with standard HDDs in it has much lower performances than my desktop with the archive HDD. Be it read or sustained write, the Ultra 2 NAS with 7500 rpm disks is far behind. I took time to test it. If we stay to its dedicated use, it just beats any current disk price per performance wise.
So I don't think I will be disappointed with its performances in a RAID 1 setup used the way I do: backups + streaming
The main concern is really about the maintenance tasks, mainly routine snapshots and the deadly scrub.
Well how do the specs of your PC compare with that of the Ultra 2? Also, what model disks are you using in the Ultra 2?
Certainly when looking at current generation disks, the Archive disks are pretty slow.
Well I would probably schedule 3-4 scrubs a year max.
Desktop is an i5 clocked at 3.4 GHz with 32 Gb of RAM, so yes, it is faster than my Ultra 2 NAS.
The disks in the NAS are Seagate NAS HDD 4TB - ST4000VN000 (2x): 64Mb cache and 5900 rpm
The write speed of 2 to 16Gb files is very poor from desktop to the NAS (75 MB /sec), compared from desktop disk 1 to desktop Archive HDD (150 MB /s). NAS is using a full 1Gb LAN connection
Scrubs are scheduled every 3 months
Not sure if this is useful to you: http://www.storagereview.com/seagate_archive_hdd_review_8tb
They put two drives in a synology DS1815+ and measured read/write and rebuild times for a RAID-1 array.
Building the array took ~57 hours (~3x longer than a PMR drive). Large backups averaged about 30 MB/s. Small incremental backups ran about the same speed as a PMR disk.
I think if I were doing this I'd use jbod, and set up incremental backup jobs to update the second volume from the first. That reduces the stress of the long resync times, and gives you the option of switching to the second volume when the first had poor performance. If you go that route, stagger the maintenance schedules.
I was aware of that review before buying my first unit of these disks, 6 months ago.
JBOD is certainly an alternative, but read speeds could suffer compared to a RAID 1 right?
Also, I am more concerned by this, from Netgear kb:
[Quote]
Scrubbing checks all stored data for consistency, and therefore will also discover potential errors in files that have not been accessed in a long time. If multiple hard discs are used in a RAID Array Level 1 or greater then scrubbing will also correct any errors found automatically. Scrubbing does not delete any data on the device.
[/Quote]
With jbod, there is no real need to run a scrub. Better run an md5 sum check on a regular base.
I fully understand the limitations + the risk of a failure when rebuilding the array with too large disks
Currently I run in a 4 disks mirrored backup scheme: NAS RAID 1 mirroring 1 Archive 8TB HDD on the PC via wired LAN. The 4th disk is an offsite backup done every week. All data has md5 sum checks that can be verified in the worst case to chose the right file to preserve in case of a failure or mismatch.
I'd prefer a regular disk, however, at 8TB, these are out of price
In reality, I am even waiting for the upcoming 10Tb disks Seagate promised
NETGEAR Academy
Boost your skills with the Netgear Academy - Get trained, certified and stay ahead with the latest Netgear technology!
Join Us!
