Upgrading Firmware Of Mellanox RoCE Cards for Final Windows Server 2012 RDMA Testing

Upgrading Mellanox Firmware

As we are preparing to roll out Windows Server 2012 R2 we are also updating the Mellanox cards we have. At the moment of writing the final driver & firmware for Windows Server 2012 R2 isn’t out yet, but let’s take a look at the process so you’re ready for prime time. If you need the latest public Mellanox driver for Windows Server 2012 R2 it’s here. Installing the driver is a straight forward process (upgrading servers with Mellanox drivers in place has been an issue however).

Mellanox provides good documentation on their site (http://www.mellanox.com/page/firmware_HCA_FW_identificationhttp://www.mellanox.com/page/firmware_NIC_FW_update) but for Mellanox newbies & many Windows server admins the process might be a bit more hands on than via a single installer they are used to.

What do you need?

The Windows Mellanox Firmware Tools (WinMFT). This gives you all the tools you need to get the job done.

It helps us with two things: find out Card ID and using that we can determine the PSID (Board ID) which tells us what firmware we need to down load.

The Win MFT tools are also used to burn the firmware.

Practical Tip 1: I have found that it pays to launch the installers Mellanox provides from an elevated command prompt as other wise UAC might trip up some clean finalization of a launched msi. The driver installer is more sensitive to this that the firmware installer.

Practical Tip 2: I you have OEM Mellanox cards from DELL/ HP/IBM … and they haven’t released the new firmware yet you can always burn your own. Please find the instructions here.

Walkthrough

I have a Windows Server 2012 R2 RTM running and I already installed the latest beta drivers I could find on the Mellanox site. But I’m a firmware version behind. So let’s fix this.

image

I put all the files I need in one handy spot

image

I launch an elevated command prompt

image

And from there I lauch the WinMFT installer

image

Just follow the instructions. image

image

image

image

image

Now you’re ready to determine the Device ID of your Mellanox card. From that same elevated command prompt navigate to C:Program FilesMellanoxWinMFT and run mst status

image

Grab the Device ID (marked in green) and execute following command:

flint -d /dev/mst/mt4099_pci_cr0 query

image

The Board ID (marked in yellow) is actually the PSID (more information here) will tell you what firmware to download from the Mellanox site). By the way, note this also tells you the current firmware.

You download the firmware from http://www.mellanox.com/page/firmware_download by selecting the card you have. In my case a ConnectX®-3 EN PCI-Ex Network Interface Cards (Ethernet Only NICs) and is use the Board ID to find my download.

image

All that’s left to do is burn the firmware image by executing the following command:

flint -d /dev/mst/mt4099_pci_cr0  -i C:SysAdminMellanoxFirmwarefw-ConnectX3-rel-2_30_3000-MCX312A-XCB_A2-A6-3.4.142_EN.bin burn

This requires you to confirm by typing in “y” and you can follow the process via a counter.image

When done you’ll need to reboot the server I order for the new firmware to actually get used. You can verify success by running the command again or by checking the information tab of you cards configuration settings. As you can see we’re running 2.30.3000 now.

image

So here you go. You might need to do this again after October 18th 2013 but you’re ready for now and all the testing you do is on the latest version of both the driver and the firmware. Happy testing!

ODX Speed Up VHDX Creation Times On Windows Server 2012 (R2)

Some technlogies you just need to see in action instead of reading about it. I have posted a video on Vimeo that shows ODX in action on Windows Server 2012 R2 and a DELL Compellent SAN running Storage Center 6.3.10 firmware that supports UNMAP & ODX. Watch the video here or on Vimeo itself for a better experience. It’s a rerun of the demo scripts used in my TechNet Belux Live Meeting of this week.

We demonstrate the amazing speeds at which we can create VHDX files on both a traditional clustered disk and a Cluster Shared Volume. If you have ever tried to create a lot of fixed VHD/VHDX files, especially larger one, then you really need to check out ODX and its potential. If you have a SAN or think about acquiring one make sure you get this feature and be sure that it works as advertised.

I hope you enjoy it and inspires you to look where you can leverage this technology in your own environments.

Mind the UNMAP Impact On Performance In Certain Scenarios

The Problem

Recently we’ve been trouble shooting some weird SQL Server to file backup issues. They started failing on the clock at 06:00 AM. We checked the NICs, the switches, the drivers, the LUNs, HBAs, … but it was all well. We considered over stressed buffers as the root cause or spanning tree issues but the clock steadiness of it all was weird. We tried playing with some time out parameters but with little to no avail. Until the moment it hit me, the file deletions that clean up the old backups!We had UNMAP enabled recently on the SAN.

Take a look at the screenshot below an note the deletion times underlined in red. That’s with UNMAP enabled. Above is with UNMAP disabled. The Backup jobs failed waiting for the deletion process.

image

This is a no issues if your backup target is running something prior to Windows Server 2012. if not, UNMAP is disabled by default. I know about the potential performance impact of UNMAP when deleting or more larger files due to the space reclamation kicking in. This is described here Plan and Deploy Thin Provisioning under the heading “Consider space reclamation and potential performance impact”. But as I’m quite used to talking about many, many terabytes of data I kind of forget to think of 500 to 600GB of files as “big” Embarrassed smile. But it seemed to a suspect so we tested certain scenarios and bingo!

Solutions

  1. Disable the file-delete notification that triggers real-time space reclamation. Find the following value HKEY_LOCAL_MACHINESYSTEMCurrentControlSetControlFileSystemDisableDeleteNotification and set it to 1.

    Note that: This setting is host wide, so for all LUNs. Perhaps that server has many other roles or needs to server that could benefit from UNMAP. If not this is not an issue.  It is however very efficient in avoiding issues. You can still use the Defragment and Optimize Drives tool to perform space reclamation on-demand or on a scheduled basis.

  2. Create LUNs that will have high deltas in a short time frame as fully provisioned LUNs (aka thick LUNs). As you do this per LUN and not on the host it allows for more fine grained actions than disabling UNMAP.  It makes no sense to have UNMAP do it’s work to reclaim the free space that deleting data created when you’ll just be filling up that space again in the next 24 hours in an endless cycle. Backup targets are a perfect example of this. This avoid the entire UNMAP cycle and you won’t mind as it doesn’t make much sense and fixes you issue. The drawback is you can’t do this for an existing volumes. So it has some overhead & downtime involved depending on the SAN solution you use. It also means that you have to convince you storage admins to give you fully provisioned LUNs, which might or might not be easy depending on how things are organized.

Conclusion

UNMAP has many benefits both in the physical and virtual layer. As with all technologies you have to understand its capabilities, requirements, benefits and draw backs. Without this you might run into trouble.

Windows Server 2012 R2 & Windows 8 .1 RTM Available on TechNet & MSDN Today

The Windows 8.1 (Pro) and Windows Server 2012 R2 RTM builds are available right now to the IT Pro & Developers communities via MSDN and TechNet subscriptions. Windows 8.1 RTM Enterprise edition will be available later this month.

Microsoft has decided to get the bits into our hands earlier due to our feedback that we need them to prepare for roll outs & test applications. This is great news and I’m happy Microsoft acted on the feedback they got from us.

So I’m already downloading …

image

So it seems the next weeks will get even busier than we had bargained for. Happy testing and remember, we weren’t born to follow Winking smile

See Microsoft’s announcement here