Anti Virus & Hyper-V Reloaded

The anti virus industry is both a blessing and a curse.  They protect us from a whole lot of security threats and at the same time they make us pay dearly for their mistakes or failures. Apart from those issues themselves this is aggravated that management does not see the protection it provides on a daily basis. Management only notices anti virus when things go wrong, when they lose productivity and money. And frankly when you consider scenarios like this one …

Hi boss, yes, I know we spent a 1.5 million Euros on our virtualization projects and it’s fully redundant to protect our livelihood. Unfortunately the anti virus product crashed the clusters so we’re out of business for the next 24 hours, at least.

… I can’t blame them for being a bit grumpy about it.

Recently some colleagues & partners in IT got bitten once again by McAfee with one of there patches (8.8 Patch 1 and 8.7 Patch 5). These have caused a lot of BSOD reports and they put the CSVs on Hyper-V clusters into redirected mode (https://kc.mcafee.com/corporate/index?page=content&id=KB73596). Sigh. As you can read here for the redirected mode issue they are telling us Microsoft will have to provide a hotfix. Now all anti virus vendors have their issue but McAfee has had too many issues for to long now.  I had hoped that Intel buying them would have helped with quality assurance but it clearly did not. This only makes me hope that whatever protection against malware is going to built into the hardware will be of a lot better quality as we don’t need our hardware destroying our servers and client devices. We’re also no very happy with the prospect or rolling out firmware & BIOS updates at the rate and with the risk of current anti virus products.

Aidan Finn has written before about the balance between risk & high availability when it comes to putting anti virus on Hyper-V cluster hosts and I concur with his approach:

  • When you do it pay attention to the exclusion & configuration requirements
  • Manage those host very carefully, don’t slap on just any update/patches and this includes anti virus products of cause

I’m have a Masters in biology from they days before I went head over heals into the IT business. From that background I’ve taken my approach to defending against malware. You have to make a judgment call, weighing all the options with their pros and cons. Compare this to vaccines/inoculations to protect the majority of your population. You don’t have to get a 100% complete coverage to be successful in containing an outbreak. Just a sufficiently large enough part including your most vulnerable and most at risk population. Excluding the Hyper-V hosts from mandatory anti virus fits this bill. Will you have 100% success, always? Forget it. There is no such thing.

Video Interview on CSV & Storage Design by Carsten Rachfahl

I already mentioned that during the Experts2Experts Virtualization Conference I met a lot of great people and I presented on High Performance & High Availability networking for Hyper-V clusters (10Gbps goodness). Some of the people I met I already knew from the on line community and others were unknown to me until that event. Among the attendees we found some of the usual virtualization suspects in our community like Aidan Finn, Jeff Wouters, Carsten Rachfahl, Ronnie Isherwood.

Now Carsten Rachfahl is a MVP in Virtual Machine expertise but he’s also a dynamic entrepreneur who shows a lot of initiative. Using social media he is really making in effort to get people & customers to notice important snippets of information by providing easy and fast access to them. He’s very active as a speaker, on Twitter and on his blogs. On top of that he does podcasts and video interviews. For Hyper-V information go to http://www.hyper-v-server.de/  which you can also use  as an entry point for his other sites focusing on several aspects of IT in the Microsoft sphere in Germany. Like cloud computing & Licensing. There you’ll also find the videos of interviews on these subjects. It’s quite an impressive endeavor.

Carsten took the opportunity to make some videos with all the above suspects on various subject and he recently released our interview. 2011-12-01-didier-interview

In this video we continued the discussion that Aidan started on CSV and we briefly touched on a subject you could make hour long documentaries about: storage options in Windows Hyper-V now and in the years to come. Enjoy!

Direct Connect iSCSI Storage To Hyper-V Guest Benefits From VMQ & Jumbo Frames

As I was preparing a presentation on Hyper-V cluster high available & high performance networking by, you guessed it, presenting it. During that presentation I mentioned Jumbo Frames & VMQ (VMDq in Intel speak)  for the virtual machine, Live Migration and CSV network. Jumbo frames are rather well know nowadays but VMQ is still something people have read about, at best have tinkered with, but no many are using it in production.

One of the reason for this that it isn’t explained and documented very well. You can find some decent explanation on what it is and does for you but that’s about it. The implementation information is woefully inadequate and, as with many advanced network features, there are many hiccups and intricacies. But that’s a subject for another blog post. I need some more input from Intel and or MSFT before I can finish that one.

Someone stated/asked that they knew that Jumbo frames are good for throughput on iSCSI networks and as such would also be beneficial to iSCSI networks provided to the virtual machines. But how about VMQ? Does that do anything at all for IP based storage. Yes it does. As a matter of fact It’s highly recommend by MSFT IT in one of their TechEd 2010 USA presentations on Hyper-V and storage.

So yes enable VMQ on both NIC ports used for iSCSI to the guest. Ideally these are two dedicated NICs connected to two separate switches to avoid a single point of failure. You do not need to team these on the host or have Multiple Path I/O (MPIO) running for this mat the parent level. The MPIO part is done in the virtual machines guests themselves as that’s where the iSCSI initiator lives with direct connect. And to address the question that followed, you can also use Multiple Connections per Session (MCS) in the guest if your storage device supports this but I must admit I have not seen this used in the wild. And then, finally coming to the point, both MPIO and MCS work transparently with Jumbo Frames and VMQ. So you’re good to go Smile

Assigning Large Memory To Virtual Machine Fails: Event ID 3320 & 3050

We had a kind reminder recently that we shouldn’t forget to complete all steps in a Hyper-V cluster node upgrade process. The proof of a plan lies in the execution Smile. We needed to configure a virtual machine with a whooping 50GB of memory for an experiment. No sweat, we have plenty of memory in those new cluster nodes. But when trying to do so it failed with a rather obscure error in System Center Virtual Machine Manager 2008 R2

Error (12711)

VMM cannot complete the WMI operation on server hypervhost01.lab.test because of error: [MSCluster_Resource.Name="Virtual Machine MYSERVER"] The group or resource is not in the correct state to perform the requested operation.

(The group or resource is not in the correct state to perform the requested operation (0x139F))

Recommended Action

Resolve the issue and then try the operation again.

image

One option we considered was that SCVMM2008R2 didn’t want to assign that much memory as one of the old host was still a member of the cluster and “only” has 48GB of RAM. But nothing that advanced was going on here. Looking at the logs found the culprit pretty fast: lack of disk space.

We saw following errors in the Microsoft-Windows-Hyper-V-Worker-Admin event log:

Log Name:      Microsoft-Windows-Hyper-V-Worker-Admin
Source:        Microsoft-Windows-Hyper-V-Worker
Date:          17/08/2011 10:30:36
Event ID:      3050
Task Category: None
Level:         Error
Keywords:     
User:          NETWORK SERVICE
Computer:      hypervhost01.lab.test
Description:
‘MYSERVER’ could not initialize memory: There is not enough space on the disk. (0x80070070). (Virtual machine ID DEDEFFD1-7A32-4654-835D-ACE32EEB60EE)

Log Name:      Microsoft-Windows-Hyper-V-Worker-Admin
Source:        Microsoft-Windows-Hyper-V-Worker
Date:          17/08/2011 10:30:36
Event ID:      3320
Task Category: None
Level:         Error
Keywords:     
User:          NETWORK SERVICE
Computer:      hypervhost01.lab.test
Description:
‘MYSERVER’ failed to create memory contents file ‘C:ClusterStorageVolume1MYSERVERVirtual MachinesDEDEFFD1-7A32-4654-835D-ACE32EEB60EEDEDEFFD1-7A32-4654-835D-ACE32EEB60EE.bin’ of size 50003 MB. (Virtual machine ID DEDEFFD1-7A32-4654-835D-ACE32EEB60EE)

Sure enough a smaller amount of memory, 40GB, less than the remaining disk space on the CSV did work. That made me remember we still needed to expand the LUNS on the SAN to provide for the storage space to store the large BIN files associated with these kinds of large memory configurations. Can you say "luxury problems"? The BIN file contains the memory of a virtual machine or snapshot that is in a saved state. Now you need to know that the BIN file actually requires the same disk space as the amount of physical memory assigned to a virtual machine. That means it can require a lot of room. Under "normal" conditions these don’t get this big and we provide a reasonable buffer of free space on the LUNS anyway for performance reasons, growth etc. But this was a bit more than that buffer could master.

As it was stated in the planning that we needed to expand the LUNS a bit to be able to deal with this kind of memory hogs this meant that the storage to do so was available and the LUN wasn’t maxed out yet. If not, we would have been in a bit of a pickle.

So there you go a real life example of what Aidan Finn warns about when using dynamic memory. Also see KB 2504962 “Dynamic Memory allocation in a Virtual Machine does not change although there is available memory on the host” which discusses the scenario where dynamic memory allocation seems not to work due to lack of disk space. Don’t forget about your disk space requirements for the bin files when using virtual machines with this much memory assigned. They tend to consume considerable chunks of your storage space. And even if you don’t forget about it in your planning, please don’t forget the execute every step of the plan Winking smile