Short questions

[yusuke shimizu]

Hello.

This is Yusuke from Redmond.

May I ask some questions?

Q1

What are the enhancement points of Health Services?

Q2

Will Health Services be able to be operated from Server Manager?

Q3

What kind of information will we be able to collect using Time Series or Alerts?

Q4

Are there any requirements for PMEM, such as endurance?

Q5

Will WS2016 support SATA controller for S2D?

Q6

Do you expect to use M.2 devices as data store?

Q7

What is Scoped Spaces?

Q8

What is Marginal Drive Handling?

Q9

What is an expected scenario for Storage Replica DR Simulation?

Q10

Will MSDTC support S2D and CSV?

Thanks

[#2]

Q10

Will MSDTC support S2D and CSV?

[RobHind]: Yes, MSDTC will be supported on S2D and CSV - but still requires DHCP capable network in the Fall release.

[#3]

Q1

What are the enhancement points of Health Services?

Answer:  In the initial release of the Health Service in Windows Server 2016, it was focused on Storage.  In the Windows Server, version 1709 release it will be expanded to covering Compute as well.  And we will continue to invest and expand the scenarios...

Thanks!
Elden

[#4]

Q5
Will WS2016 support SATA controller for S2D?
A5
Yes, but S2D requires SES so there needs to be a SES controller.
Q7
What is Scoped Spaces?
A5
Scoped spaces is a new functionality in the 1709 release that allows specifying which nodes of an S2D system that the data will be put on when creating a new volume. If you have 8 nodes, you can scope some volumes to 4, and other volumes to other sets of 4. This means that if you have more than 2 failures across nodes it will limit the number of volumes effected.

[#5]

Hello Rob,

Thank you for the answer.

Does "Fall release" mean RS3?

[#6]

For Q4

Are there any requirements for PMEM, such as endurance?

There are no Windows Server specific requirements for PMEM as currently available technology (NVDIMM-N) has a high enough req and do not need anything more.

Recently we published a blog on endurance for more information

https://blogs.technet.microsoft.com/filecab/2017/08/11/understanding-dwpd-tbw/

[#7]

Yes. "Fall Release" = RS3 = 1709.

[#8]

Hello Steven,

Thanks for the answers.

Regarding A5, will WS2016RTM support SES controller for S2D?

Or, is RS3 required?

[#9]

Q9

What is an expected scenario for Storage Replica DR Simulation?

A:  Windows Server, version 1709 gives you the ability to test failover, and mount read-write snapshot on destination which is useful to view contents, test a VM, copy data off, run backups, etc...

Thanks!
Elden

[#10]

Dear All,

I really appreciate all your help!

[#11]

Q1
What are the enhancement points of Health Services?

[Rob for Cosmos]:

The big two enhancements are:

(a) Better defenses against marginal drives

(b) Improved performance metrics for storage and now for compute!

[#12]

Q2
Will Health Services be able to be operated from Server Manager?

[Rob for Cosmos]:
We are working on a graphical experience for the Health Service – stay tuned!

[#13]

Q3
What kind of information will we be able to collect using Time Series or Alerts?

[Rob for Cosmos]:
Timeseries collects over 50 total series across Cluster, Server, VM, VHD, Volume, and Drive.
This includes CPU, Memory, storage IOPS, IO latency, IO throughput, and even network and RDMA activity.
The exact series will be published shortly, we’re writing the documentation now!

[#14]

Q6
Do you expect to use M.2 devices as data store?

[Rob for Cosmos]:
There is no reason (that I can think of) preventing M.2 devices being used as cache or capacity for the Storage Spaces Direct pool.
We don’t really care about the physical form factor, as long as it speaks NVMe.
However, a consideration would be serviceability – it’s internal and therefore hard to access, and it’s generally not safe to replace without powering down the server.
AFAIK, the main appeal of M.2 is the small size, e.g. for laptops. For servers, I don’t see a clear benefit - but it should work.

[#15]

Q8
What is Marginal Drive Handling?

[Rob for Cosmos]:
Documentation on this will be coming soon. In general, we are hardening Storage Spaces Direct with defenses against drives which exhibit too many bad blocks, too many IO errors or retries, or (in next year’s spring update) outlier latency.
After certain (configurable) thresholds are exceeded, we mark the drive as Warning and suggest that the user should consider replacing it.

[#16]

Hello Rob,

I really appreciate your follow-up.