Hi Toasters,
Curious if anybody has experience with an all flash FAS deployment? We currently utilize Flash Cache and Flash Pool backed by SAS 15K and 10K drives but are considering an all flash FAS solution for our next system.
What prompted you to go with an all flash config versus leveraging flash as a cache tier? What size SSD's are you using?
Not interested in an E-series array as we depend upon the features of ONTAP. Also not interested in fronting an E-series array with an ONTAP system. I am aware of the 96 disk per stack limitation that goes along with FAS and flash.
Cheers! -Phil
Also interested in this topic. I attended NetApp Insight 2015 and spoke with a few NetApp engineers about "speeding up" OLTP workload, 100% response was "go flash pool or all flash". Hopefully soon I'll an access to a lab with FAS system with both hybrid and all flash aggregates.
Cheers, Vladimir
On Thu, Mar 19, 2015 at 11:21 PM, Philbert Rupkins philbertrupkins@gmail.com wrote:
Hi Toasters,
Curious if anybody has experience with an all flash FAS deployment? We currently utilize Flash Cache and Flash Pool backed by SAS 15K and 10K drives but are considering an all flash FAS solution for our next system.
What prompted you to go with an all flash config versus leveraging flash as a cache tier? What size SSD's are you using?
Not interested in an E-series array as we depend upon the features of ONTAP. Also not interested in fronting an E-series array with an ONTAP system. I am aware of the 96 disk per stack limitation that goes along with FAS and flash.
Cheers! -Phil
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
I have to say - I've been bouncing around the concept for a while - whether going for an AFA is 'worth it'. I actually think that Filers ... are less so than some options. For writes, between WAFL, RAID-DP and NVRAM .... it's actually not all that often you need fast sustained throughput.
For reads... well, yes, the read latency is always a factor. But it remains a question - how much true random read do you do? Between 'stats show -p flexscale-access' and 'priv set diag; stats show wafl:wafl' which will show you where the reads are coming from.
For most of my filer heads, I'm getting pretty high read cache hit rates - 90%+ (between 100G of RAM and 2TB of PAM) leads me to conclude that I'd be wasting my money going all flash.
So it's really a workload question. If you've a lot of write IO, then more controllers and spindles will be as good. (short stroking if necessary).
The place where all flash really delivers is for true random-read workloads. If you have one of those you'll be paying the seek latency for most IOs, at which point having flash will pay dividends. You'll be able to see this, because your read cache hit rate will be really low. And your service users are complaining about read latency being too high.
I've not seen the need yet, but then I don't have any workloads like that. For most, a bit of cache (RAM + PAM) - and maybe for extremes going for hybrid aggregates - gives me the better cost-benefit ratio.
On 20 March 2015 at 20:57, Momonth momonth@gmail.com wrote:
Also interested in this topic. I attended NetApp Insight 2015 and spoke with a few NetApp engineers about "speeding up" OLTP workload, 100% response was "go flash pool or all flash". Hopefully soon I'll an access to a lab with FAS system with both hybrid and all flash aggregates.
Cheers, Vladimir
On Thu, Mar 19, 2015 at 11:21 PM, Philbert Rupkins philbertrupkins@gmail.com wrote:
Hi Toasters,
Curious if anybody has experience with an all flash FAS deployment? We currently utilize Flash Cache and Flash Pool backed by SAS 15K and 10K drives but are considering an all flash FAS solution for our next system.
What prompted you to go with an all flash config versus leveraging flash
as
a cache tier? What size SSD's are you using?
Not interested in an E-series array as we depend upon the features of
ONTAP.
Also not interested in fronting an E-series array with an ONTAP system.
I
am aware of the 96 disk per stack limitation that goes along with FAS and flash.
Cheers! -Phil
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
On Fri, Mar 20, 2015 at 10:10:04PM +0000, Edward Rolison wrote:
I have to say - I've been bouncing around the concept for a while - whether going for an AFA is 'worth it'. I actually think that Filers ... are less so than some options. For writes, between WAFL, RAID-DP and NVRAM .... it's actually not all that often you need fast sustained throughput.
For reads... well, yes, the read latency is always a factor. But it remains a question - how much true random read do you do? Between 'stats show -p flexscale-access' and 'priv set diag; stats show wafl:wafl' which will show you where the reads are coming from.
For most of my filer heads, I'm getting pretty high read cache hit rates - 90%+ (between 100G of RAM and 2TB of PAM) leads me to conclude that I'd be wasting my money going all flash.
So it's really a workload question. If you've a lot of write IO, then more controllers and spindles will be as good. (short stroking if necessary).
The place where all flash really delivers is for true random-read workloads. If you have one of those you'll be paying the seek latency for most IOs, at which point having flash will pay dividends. You'll be able to see this, because your read cache hit rate will be really low. And your service users are complaining about read latency being too high.
I've not seen the need yet, but then I don't have any workloads like that. For most, a bit of cache (RAM + PAM) - and maybe for extremes going for hybrid aggregates - gives me the better cost-benefit ratio.
We've looked at AFA's a few times in the past as well, and as cool as it would be to have one, I really haven't found the business need for it. Maybe if we did large-scale VDI?
Hybrid approach works really well for us for our typical mixed-used / random read heavy workloads. Using IBM V7000+EasyTier+compression with very good results and easy to just add in more SSD's. Lets me size much larger than what I could cost-effectively get from PAM. I think FAS may have a hybrid option now though?
Still use plenty of FAS+PAM for NAS though. V7K Unified has a ways to go....
Ray
-
Edward Rolison -
Momonth -
Philbert Rupkins -
Ray Van Dolson