-
RAID level reliability
-
[Vadim Carter] “ZFS storage pools can span multiple vdevs (virtual devices)”
Vadim, I was talking about how the width of a stripe affects performance. Can’t improve performance without widening a stripe one way or the other, correct? I.e. even in RAIDZ, you’d have to re-stripe the array to improve performance. (At least that’s my understanding.)
-
[Alex Gerulaitis] “Vadim, I was talking about how the width of a stripe affects performance. Can’t improve performance without widening a stripe one way or the other, correct? I.e. even in RAIDZ, you’d have to re-stripe the array to improve performance. (At least that’s my understanding.)”
There are a few ways to improve performance, and, you are correct, Alex, by increasing the stripe width or, simply put, striping across more disk spindles thus taking advantage of parallelism. When you expand a ZFS pool, ZFS uses dynamic striping in order to maximize throughput and attempts to include all devices in order to balance it.
Below is a quote from Oracle:
“ZFS dynamically stripes data across all top-level virtual devices. The decision about where to place data is done at write time, so no fixed-width stripes are created at allocation time.
When new virtual devices are added to a pool, ZFS gradually allocates data to the new device in order to maintain performance and disk space allocation policies. Each virtual device can also be a mirror or a RAID-Z device that contains other disk devices or files. This configuration gives you flexibility in controlling the fault characteristics of your pool. For example, you could create the following configurations out of four disks:
-Four disks using dynamic striping
-One four-way RAID-Z configuration
-Two two-way mirrors using dynamic striping
Although ZFS supports combining different types of virtual devices within the same pool, avoid this practice. For example, you can create a pool with a two-way mirror and a three-way RAID-Z configuration. However, your fault tolerance is as good as your worst virtual device, RAID-Z in this case. A best practice is to use top-level virtual devices of the same type with the same redundancy level in each device.”
Lucid Technology, Inc. / 801 West Bay Dr. Suite 465 / Largo, FL 33770
“Enterprise Data Storage for Everyone!”
Ph.: 727-487-2430
https://www.lucidti.com -
At our house , I have tested a lot on a “simple” 12 bay shelf, with a Dell H800 hw raid card (Level 5, 6 10, 50 and 5 + 5 mirror) , and now i have used LSI 9211-8i to also test ZFS and i have to say, im deeply impressed.
There is a lot of hard geeking to do, but if u sett it up right, you will gain a lot of performance, and zfs has some other cool features. like nfs sharing built in the file system, and compression, and as mentioned checksums, on the comercial front you have “NexentraStor” (OpenIndiana / SUN) and “ixsystems” (Freenas / FreeBSD) , you also have, napp-it a pluggin for a lot of systems, but that runs best on Omni-OS
I think i have used over a year now for testing, and we landed on FreeBSD 9.1 with Freenas (with comercial support from ixsystems)
Some interesting links that i have picked up ower the year.
https://hansdeleenheer.blogspot.no/2012/07/nexenta-scale-and-cluster.html
https://calomel.org/zfs_raid_speed_capacity.html
https://catn.com/2012/05/11/openindiana-vs-nexentastor-vs-freenas/
https://www.zfsbuild.com/2013/01/25/zfsbuild2012-nexenta-vs-freenas-vs-zfsguru/Best “old” video explaining zfs
https://www.youtube.com/watch?v=3-KesLwobps -
Thanks for your post, Andreas. It is nice to get an independent endorsement from a person using RAIDz/zfs in production. While there are still many users who are completely unaware of the benefits zfs-based storage brings to the table, I believe the tide has turned and we’ll see more people jumping aboard.
Lucid Technology, Inc. / 801 West Bay Dr. Suite 465 / Largo, FL 33770
“Enterprise Data Storage for Everyone!”
Ph.: 727-487-2430
https://www.lucidti.com -
[Andreas Aanerud] “There is a lot of hard geeking to do, but if u sett it up right, you will gain a lot of performance”
I would love to hear about that “lot of hard geeking”, and why it’s necessary, how users deal with and react to the lack of online array expansion, the fact that ZFS only applies to NAS boxes (not DAS or SAN storage), and about ZFS performance benefits.
Yet since this is a thread dedicated purely to RAID level reliability, then perhaps we could fork ZFS-related discussions to a different thread on Cow’s NAS forum?
Reply to this Discussion! Login or Sign Up