Hello All,

I'm facing a problem with S2D where the performance is slow. Please help.

We have Intel x722 10GB with RDMA running. 4 Intel P4600 1.6TB Mainstream NVMe PCIe 3.0 x4 Hot Swap SSD 8 8TB HDD Per Node

Command Line: diskspd -b64K -d60 -o8 -t8 -h -r -w70 -L -c20G C:\ClusterStorage\Hyper V Machines\Test\io.dat

Input parameters:

timespan:   1
-------------
duration: 60s
warm up time: 5s
cool down time: 0s
measuring latency
random seed: 0
path: 'C:\ClusterStorage\Hyper V Machines\Test\io.dat'
think time: 0ms
burst size: 0
software cache disabled
hardware write cache disabled, writethrough on
performing mix test (read/write ratio: 30/70)
block size: 65536
using random I/O (alignment: 65536)
number of outstanding I/O operations: 8
thread stride size: 0
threads per file: 8
using I/O Completion Ports
IO priority: normal

System information:

computer name: RTPNTSSSD01
start time: 2019/03/20 14:08:19 UTC

Results for timespan 1:
*******************************************************************************

actual test time:60.00s
thread count:8
proc count:32

CPU |  Usage |  User  |  Kernel |  Idle
-------------------------------------------
   0|  43.80%|   6.90%|   36.90%|  56.20%
   1|  18.44%|   4.27%|   14.17%|  81.56%
   2|  40.23%|   7.47%|   32.76%|  59.77%
   3|  20.57%|   5.44%|   15.13%|  79.43%
   4|  31.02%|   6.28%|   24.74%|  68.98%
   5|  31.77%|   7.99%|   23.78%|  68.23%
   6|  39.66%|   9.56%|   30.10%|  60.34%
   7|  18.31%|   4.82%|   13.49%|  81.69%
   8|  48.93%|  12.73%|   36.20%|  51.07%
   9|  26.20%|   5.26%|   20.94%|  73.80%
  10|  51.12%|  10.08%|   41.04%|  48.88%
  11|  18.93%|   3.96%|   14.97%|  81.07%
  12|  41.43%|   7.47%|   33.96%|  58.57%
  13|  28.78%|   7.86%|   20.91%|  71.22%
  14|  38.80%|   8.23%|   30.57%|  61.20%
  15|  32.11%|   5.55%|   26.56%|  67.89%
  16|  42.11%|  14.71%|   27.40%|  57.89%
  17|  32.92%|  18.33%|   14.58%|  67.08%
  18|  33.39%|   9.61%|   23.78%|  66.61%
  19|  33.59%|  23.39%|   10.21%|  66.41%
  20|  41.04%|  11.33%|   29.71%|  58.96%
  21|  25.18%|  13.41%|   11.77%|  74.82%
  22|  40.76%|  13.96%|   26.80%|  59.24%
  23|  15.94%|   6.04%|    9.90%|  84.06%
  24|  28.02%|   8.80%|   19.22%|  71.98%
  25|  31.88%|  13.85%|   18.02%|  68.13%
  26|  28.67%|  10.18%|   18.49%|  71.33%
  27|  32.01%|  19.32%|   12.68%|  67.99%
  28|  50.10%|  15.96%|   34.14%|  49.90%
  29|  15.10%|   6.80%|    8.31%|  84.90%
  30|  35.44%|  14.04%|   21.41%|  64.56%
  31|  18.49%|   6.30%|   12.19%|  81.51%
-------------------------------------------
avg.|  32.34%|  10.00%|   22.34%|  67.66%

Total IO
thread |       bytes     |     I/Os     |    MiB/s   |  I/O per s |  AvgLat  | LatStdDev |  file
-----------------------------------------------------------------------------------------------------
     0 |       153944064 |         2349 |       2.45 |      39.15 |  204.389 |    79.250 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     1 |       151060480 |         2305 |       2.40 |      38.42 |  208.384 |    72.289 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     2 |       149749760 |         2285 |       2.38 |      38.08 |  210.314 |    78.447 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     3 |       153550848 |         2343 |       2.44 |      39.05 |  205.052 |    77.734 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     4 |       152502272 |         2327 |       2.42 |      38.78 |  206.276 |    77.226 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     5 |       151977984 |         2319 |       2.42 |      38.65 |  207.071 |    73.804 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     6 |       151322624 |         2309 |       2.41 |      38.48 |  207.967 |    75.746 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     7 |       155058176 |         2366 |       2.46 |      39.43 |  203.011 |    78.943 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
-----------------------------------------------------------------------------------------------------
total:        1219166208 |        18603 |      19.38 |     310.05 |  206.534 |    76.760

Read IO
thread |       bytes     |     I/Os     |    MiB/s   |  I/O per s |  AvgLat  | LatStdDev |  file
-----------------------------------------------------------------------------------------------------
     0 |        44695552 |          682 |       0.71 |      11.37 |  211.776 |    79.911 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     1 |        44367872 |          677 |       0.71 |      11.28 |  214.141 |    70.812 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     2 |        45088768 |          688 |       0.72 |      11.47 |  216.849 |    80.378 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     3 |        45416448 |          693 |       0.72 |      11.55 |  212.791 |    81.773 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     4 |        47448064 |          724 |       0.75 |      12.07 |  216.560 |    78.843 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     5 |        46137344 |          704 |       0.73 |      11.73 |  211.739 |    73.200 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     6 |        44826624 |          684 |       0.71 |      11.40 |  217.095 |    78.534 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     7 |        45744128 |          698 |       0.73 |      11.63 |  210.428 |    81.585 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
-----------------------------------------------------------------------------------------------------
total:         363724800 |         5550 |       5.78 |      92.50 |  213.925 |    78.269

Write IO
thread |       bytes     |     I/Os     |    MiB/s   |  I/O per s |  AvgLat  | LatStdDev |  file
-----------------------------------------------------------------------------------------------------
     0 |       109248512 |         1667 |       1.74 |      27.78 |  201.367 |    78.779 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     1 |       106692608 |         1628 |       1.70 |      27.13 |  205.990 |    72.761 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     2 |       104660992 |         1597 |       1.66 |      26.62 |  207.498 |    77.431 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     3 |       108134400 |         1650 |       1.72 |      27.50 |  201.801 |    75.739 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     4 |       105054208 |         1603 |       1.67 |      26.72 |  201.632 |    76.030 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     5 |       105840640 |         1615 |       1.68 |      26.92 |  205.036 |    73.974 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     6 |       106496000 |         1625 |       1.69 |      27.08 |  204.125 |    74.206 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
     7 |       109314048 |         1668 |       1.74 |      27.80 |  199.907 |    77.601 | C:\ClusterStorage\Hyper V Machines\Test\io.dat (20GiB)
-----------------------------------------------------------------------------------------------------
total:         855441408 |        13053 |      13.60 |     217.55 |  203.392 |    75.892



total:
  %-ile |  Read (ms) | Write (ms) | Total (ms)
----------------------------------------------
    min |      0.069 |      0.603 |      0.069
   25th |    163.372 |    153.395 |    156.194
   50th |    204.242 |    193.350 |    196.290
   75th |    255.872 |    242.513 |    246.704
   90th |    310.354 |    296.396 |    300.650
   95th |    354.818 |    338.732 |    344.912
   99th |    452.223 |    446.069 |    447.682
3-nines |    531.506 |    528.533 |    529.401
4-nines |    576.729 |    577.402 |    577.402
5-nines |    576.729 |    577.611 |    577.611
6-nines |    576.729 |    577.611 |    577.611
7-nines |    576.729 |    577.611 |    577.611
8-nines |    576.729 |    577.611 |    577.611
9-nines |    576.729 |    577.611 |    577.611
    max |    576.729 |    577.611 |    577.611

Hello all!

Configuration-> Failover File server cluster in S2d with 2 Dell server with 2 "internal" drive for OS, 4 X 1.4TG SSD for caching/Journal, and 8 X 4TB for Data

Currently, Server 1 is providing file access to the users.

On server 2, we got a warning of one of the 4 SSD being "dead". Contacted Dell and a replacement drive was sent BUT bigger then the original (i.e. a 1.9TB instead of 1.6TB).

Called Dell and talk to different support personal and was told it was OK to replace with the bigger one (although I still had reservation doing it).

Now I find myself wondering if all my Cluster and S2D is still correct since when I check, the "new" SSD shows up as the usage being "Automatic" instead of "Journal" as the other SSD drives in both servers.

Did I do (or not do) somethin wrong and should I worry?

Should I "remove" the new drive from the Storage pool and re-add forcing it's usage to journal?

What will happen if I have a failover (or manually switch around) between the servers in their current state?

Thanks all for any opinion, advice comment you may provide me with.

My regards

Serge Ayotte

Hi everyone,

I have a question about setting up File Services.

On my physical server (Host) I have 3 drives (let's call them C, D and E).

C drive is of course where the main Host OS installed with Hyper-V role, etc.

D drive is where I have pointed Hyper-V to save my VHD's, snapshots and whatever else my VM's would need.

E drive is where I plan to have all the Data.

The host is used for Hyper-V role only and all other services (AD, RDP, etc) will be installed on the Guest. But I don't want to have the Data inside the VM because if it crashes/fails or Guest OS simply becomes unusable then it will be a pain to extract the Data out from the VHD. So I believe it's better to have the Data on that E drive.

So the question is what's the best way to link the Data on that E drive from the Host to the Guest? Should I add is as a disk via VM settings? Or should I create a Share on the Host and then map the path to it as a network drive on the Guest? Any other options?

And I obviously want to have all the benefits of the File Server including Shadow Copies and Previous Versions, etc. So another question is where should I install File Services role for this to work best? On the Host or on the Guest? Or both?

Thanks everyone in advance.

I'm trying to use robocopy (as I have always done) to transfer data to another hard drive, but it doesn't seem to be working.  I am not getting an error, but it is not transferring data or showing that it is transferring data.  As far as I know, my path is correct!  Has Microsoft done something to Robocopy so it no longer works?

Thank you,

Todez

I set up a FTP server (Win7) over the Internet Information Services (IIS) Manager. I use it to transfer a LOG file from a controller to the Physical Path in the FTP Server. It is working but it is putting the LOG files into dated subfolders (C:\FTPClient\Year\Month\Day...). I think this is being created by the Client. Regardless, is there a setting in the FTP Server settings to redirect the LOG to only get put into the Physical Path without the subfolders?

Thanks,

It's a pretty common practice for SQL backed file stores to be formatted with 64k allocation unit size. SQL sort of has its own file system that sits ontop of NTFS anyway, so its a very large single file.

But, is using that 64k as a standard for a regular file store a good idea?

I'm trying to understand specifically what is going on at like the disk read/write head level when it comes to picking an allocation unit. Let's say I have a volume and its using 64k allocation unit size. Let's say I have an application that is copying some files, but is also writing to log files frequently. Let's also say that for diagnostic reasons, the logger opens the file (for append), writes the log message, flushes, then closes the file. It might do this many many times during the session.

Given the above, Is this a true statement?: If the logger writes 10 bytes to the file, the entire 64k allocation unit is re-written (so IO of 64 bytes written occurred). If the logger writes a 1000 byte message, again, 64k is actually written.

Or, is it smarter than that, and will only write the specific number of bytes related to the file IO operation?

Guys,

I am trying to learn icsci but i cant figure out why i should use it and how i need to install/configure it to connect to iscsi storage i  a windows 10/server2016 environment.

Anyone who knows a good tutorial? The ones i found arent helpfull enough.

Many thanks for a reply,

Hi

Current setup is a single Win2012R2 server in a failover cluster. The other Win2012R2 node has been shut down, so not much failover clustering left ;)

This server (SERVER01) runs DFS (standalone namespaces) which is accessed by the clients using the name OURDFS. So drives are mapped with \\OURDFS\Shares01\Data01,\\OURDFS\Shares02\Data01 and so on..... All data being accessed through DFS is located on file servers on the network, so it's only DFS Namespace service this server is hosting.

Now this server needs to be closed down pretty soon, so I need to move DFS to another 2012R2 server (or newer version). For reasons I don't wanna address here, we still need to be able to access the drives using the same UNC paths.

Method 1: I have thought about taking SERVER01 offline and just building a new server which I call OURDFS and configure it with the same namespaces and folders, but this probably is a point of no return as I can't fall back to the old SERVER01 if anything goes wrong as I'll be removing the OURDFS AD computer object created by the clustering service when I join the new server to the domain. So no going back from this point.

Method 2: Also thought about adding another node to the cluster (SERVER02) and then shutting down SERVER01, but this leaves me with a cluster I actually don't need anymore and I fear I can't remove the cluster without also removing the AD computer object and then messing up DFS. Probably could then fix it by just renaming the server but this brings me back to method 1 i believe.....    

So I actually don't not know which way is the better way when I wanna minimize downtime and most importantly of all my weekend-work time :)

Perhaps I'm missing a better solution or perhaps it's no biggie and I should just go ahead with one of the mentioned methods.

Also if you've got any caveats please help me out by mentioning them :)

Best

Thomas

I'm automatically archiving Windows Server Security logs and I've configured Search Indexing service to index the Properties and File Contents of the resultant .evtx files, but I'm still not able to search the contents of the files. Anyone know what I'm doing wrong? Is this just not possible, even thought it's an option in the Index settings?

Shaun

I can't find where to set auto cleanup temp file in window server 2016.  It doesn't show me storage sense under storage such I can set auto cleanup.  Is there such a feature in windows server 2016?

Thank you,

NLuu

Hello

I have been using Storage spaces for over 2 years without much of a problem until now.

I have 6 drives in the pool totaling about 10tb.  

10 days ago, I was doing some clean up and deleted 350gb of data but the pool availability did not change.

I noticed that the recycle bin was full so I figured that I need to empty it out.  Well that's where it went wrong.

Since emptying the recycle bin, the storage space drive disappeared from the windows explorer.  I can see the drives are "working" and if I go in Storage Spaces, it is showing all 6 drives are in good condition with about 70% capacity.  All options other than rename are grey out.

Now it has been 10 days now and I re-started my computer once to see whether it did any good, well no, it did not.

Does everyone knows what's wrong or should I just wait it out for the pool to sort itself out.

Please advise.

Many thanks in advance

Henry

Hi,

Is there any in built alternative to Dokan user mode file System driver, windows is providing.

Currently libfuse is part Linux OS.

Please let me know, if something similar to Dokan is already there which can be used.

Thanks,

Jamey 

Hi,

I have a WS2019 stretch cluster lab running Storage Replica Async and I have managed to break the replication, hoping someone can suggest how best to recover from a scenario like this.

It was working fine, and I actually enabled Deduplication on the cluster file server and tested that out. It appeared to be ok, but then I attempted to move the cluster group to another node and at this point Storage Replica failed with this error:

Log service encountered a corrupted metadata file.

I assume that the cluster may not have liked the fact that there were writes going on at the time the async disk replication attempted failover -- whether standard filesystem writes or Deduplication optimisation I'm not sure.

Now that it is in this state, how do I recover? If I attempt to online the disk resource it comes online for a few seconds then repeats the above error. Is there a way to repair or reset the log without starting from scratch? Or do I just need to disable replication and recreate it?

Thanks,
Simon.

Hi all,

We are migrating users to office 365 OneDrive and we need a way to automatically stop work folders and delete the data on the client.

Is there a way to do this without user interaction?

Rahamim.

Good morning, 

i created a FTP site and a folder in my WS2012.

inside of it i have several folders, i wish each of them would be visible by only 1-2 users. How can i do? if now i enter in the FTP i can access to all the folders. Should i use the NTFS permissions?

Furthermore i wish to understand how to use the "guest" account for accessing a folder.

Hu

I am looking for a solution that can encrypt data on a single Windows file share for HR information, clients are also Windows,

Are there any solutions for this

Regards Mogens

Hi, when using Get-DfsnFolderTarget remotely such as Invoke-Command -ComputerName Servername {Get-DfsnFolderTarget \\domain\dfs\share} the following error is returned:

Get-DfsnFolderTarget : Cannot get DFS folder properties on "\\domain\dfs\share"
    + CategoryInfo          : PermissionDenied: (MSFT_DfsNamespaceFolderTarget:ROOT\Microsoft\...aceFolderTarget) [Get-DfsnFolderTarget], CimException
    + FullyQualifiedErrorId : Windows System Error 5,Get-DfsnFolderTarget

Get-DfsnFolderTarget : Access to a CIM resource was not available to the client.
    + CategoryInfo          : PermissionDenied: (MSFT_DfsNamespaceFolderTarget:ROOT\Microsoft\...aceFolderTarget) [Get-DfsnFolderTarget], CimException
    + FullyQualifiedErrorId : MI RESULT 2,Get-DfsnFolderTarget

When using New-PSSession ServerName, Enter-PSSession, Get-DfsnFolderTarget \\domain\dfs\share I get the same result.

Running the command locally does return a result. It appears all DFS cmdlets give this error. Is this a double hop problem?

Hi,

anything i should know before I renaming a folder to an Administrative Share test$. permission right now are good. 

thanks

 second Dc controller is shown private network

and error in operations mater rules

and in the event viewer, these events appear:

and this link for Dcdiag, and DC Ipconfig and replication.

https://1drv.ms/f/s!Ag-u-xMhnkazgm8g2xdDHBiNyqKg