—– Original Message —–
From: "Victor Miasnikov"
Sent: Thursday, September 19, 2013 3:49 PM
Subject: Re: turn off 220V on UPS device =} file system got corrupted Re: Hyper-V 2012 Cluster / Failover – supported? – Any known issues?
A.G.> if high availability failover scenarios will work for FreeBSD VMs on Hyper-V.
A.G.>if the power plug is pulled from the Hyper-V server
A.G.>then would the FreeBSD VM failover and restart without any issues on the failover server.
Karl, are You want this behavior:==
you walk up and yank the power cord out of the back of the server the secondary mirror will take over with zero client
Karl, are You use entry level fault tolerant system ftServer 2600 by Stratus Technologies? Or analog?
If "no use" , then read some info about real Hyper-V Fault Tolerance :
. . .
From: Victor Miasnikov
Sent: Wednesday, September 18, 2013 8:46 AM
To: Abhishek Gupta (LIS); Karl Pielorz;
Subject: turn off 220V on UPS device =} file system got corrupted Re: Hyper-V 2012 Cluster / Failover – supported? – Any known issues?
K.P.> – Pulling the power on the active node hosting both VM’s (i.e. Windows
K.P.> guest, and FreeBSD guest) – this showed the remaining node trying to bring
K.P.> up the VM’s (of which Windows came up OK, and FreeBSD [file system] got corrupted).
A.G.> Yes, it should work.
A.G.>My understanding is that the failover should be agnostic to the guest OS but there could be some integration
component that we might have missed.
What _exactly_ "should work" ?
1) This issue not related Hyper-V cluster itself
!) When "Pulling the power" i.e. turn off 220V in Europa ( or 110V in USA ) on UPS device _both_ FAT on Windows and
FreeBSD [file system] got corrupted
( "Windows came up OK" look like because on this VM file system is NTFS )
K.P.> Hyper-V correctly see’s the node fail, and restarts both VM’s on the
K.P.> remaining node. Windows 7 boots fine (says it wasn’t shut down correctly –
K.P.> which is correct) – but FreeBSD doesn’t survive.
K.P.> At boot time we get a blank screen with "-" on it (i.e. the first part of
K.P.> the boot ’spinner’) – and nothing else.
K.P.> Booting to a network copy of FreeBSD and looking at the underlying virtual
K.P.> disk – it appears to be trashed. You can mount it (but it understandably
K.P.> warns it’s not clean) – however, any access leads to an instant panic (’bad
K.P.> dir ino 2 at offset 0: mangled entry’).
K.P.> Trying to run fsck against the file system throws up an impressive amounts
K.P.> of ’bad magic’ errors and ’rebuild cylinder group?’ prompts.
To Karl: I ask You about some details . . .
Are You see related e-mail?
Best regards, Victor Miasnikov
Abhishek Gupta (LIS)
Sep 18, 2013; 9:18pmRE: turn off 220V on UPS device =} file system got corrupted Re: Hyper-V 2012 Cluster
Karl is asking if high availability failover scenarios will work for FreeBSD VMs on Hyper-V. He was specifically interested in knowing if the power plug is pulled from the Hyper-V server then would the FreeBSD VM failover and restart without any issues on the failover server.
My response was that yes the above scenario should work.
Evaluating High-Availability (HA) vs. Fault Tolerant (FT) Solutions
10-06-2010 4:09 AM
High Availability Solutions
High availability solutions traditionally consist of a set of loosely coupled servers which have failover capabilities.Each system is independent and self-contained, yet the servers are health monitoring each other and in the event of a failure, applications will be restarted on a different server in the pool of the cluster.Windows Server Failover Clustering is an example of an HA solution.HA solutions provide health monitoring and fault recovery to increase the availability of applications.A good way to think of it is that if a system crashes (like the power cord was pulled), the application very quickly restarts on another system.HA systems can recover in the magnitude of seconds, and can achieve five 9’s of uptime (99.999%)… but they realistically can’t deliver zero downtime for unplanned failures.They also are flexible in that they enable recovery of any application running on any server in the cluster.
Fault Tolerant Solutions
Fault tolerant solutions traditionally consist of a pair of tightly coupled systems which provide redundancy.Generally speaking this involves running a single copy of the operating system and the application within, running consistently on two physical servers.The two systems are in lock step, so when any instruction is executed on one system, it is also executed on the secondary system.A good way to think of it is that you have two separate machines that are mirrored.In the event that the main system has a hardware failure, the secondary system takes over and there is zero downtime.
HA vs. FT
So which solution is right for you?Well, the initial and obvious conclusion most instantly come to is that ’no’ downtime is better than ’some’ downtime, so FT must be preferred over HA!Zero downtime is also the ultimate IT utopia which we all strive to achieve, which is goodness.Also FT is pretty cool from a technology perspective, so that tends to get the geek in all of us excited and interested.
However, it is important to understand they protect against different types of scenarios… and the key aspect to understand is what are the most important to you and your business requirements.It is true that FT solutions provide great resilience to hardware faults, such as if you walk up and yank the power cord out of the back of the server… the secondary mirror will take over with zero client downtime.However, remember that FT solutions are running a common operating system across those systems.In the event that there is a software fault (such as a hang or crash), both machines are affected and the entire solution goes down.There is no protection from software fault scenarios and at the same time you are doubling your hardware and maintenance costs.At the end of the day while a FT solution may promise zero downtime for unplanned failures, it is in reality only to a small set of failure conditions.With a loosely coupled HA solution such as Failover Clustering, in the event of a hang or blue screen from a buggy driver or leaky application.Then the application will failover and recover on another independent system.
I’ve got a two node server-cluster, WS 2008R2 x64, Hyper-V and CSV, Everything seems to be working fine along with live migration.
I am currently testing the functionality of the setup, he is my current layout:
Node A:VM 1
When I simulate a host failure on node A, VM 1 transfers over to Node B but reboots the virtual machine before bringing it back up.
Is this normal behavior for Clustering with CSV? I have another cluster setup in the same manner but without CSV enabled, Its been a while but I’m sure when this was tested the Virtual machine that failed over didn’t reboot.
Is this a difference between High availability and Fault tolerance?
If any of you guys can shed some like, it would of great help…
Yes, it works as expected. "High availability" does not mean "no downtime". If you want to have zero (OK, close to zero) downtime then you need either configure guest VM cluster or use your app built-in clustering features. If your app has none and it’s not cluster aware consider moving to VMware to use it’s Fault Tolerance feature (no equivalent for Hyper-V so far).
. . .
As VR38DETT says, what you are seeing is normal. Live Migration, which moves a VM from one host to another, is a planned action. You tell the clustering software to move the machine. This gives the software time to copy the contents of the memory on the currently hosting machine to the memory of the destination machine. In a failover environment, there is no time for the memory on the failed machine to get copied. Therefore, all that can happen is to start the VM with a boot to get the memory loaded into the desitination machine. That’s a pretty typical definition of high availability.
Now with 2012 coming out, there is a capability the Microsoft engineers have built in called ’replica’. It keeps a copy of the memory of a running virtual machine on another virtual machine. However, it is asynchronous, so it is not always up to the second current. But it gets much closer to what you are asking for.
Or, there are third parties, such as Stratus, that create a mirrored environment between two systems in order to keep two copies up to date. As you can imagine, there are additional costs involved in such a solution as this, so you need to make the business case for 100% availability.
And, as VR38DETT says, with additional capabiliies, like clustering the VMs at their operating system/application level can provide a different sort of [near] continuous operation. I say [near] because it is definitely dependent upon the software you are running within the VM. For example, if you are running a particular type of SQL Server, you have SQL running on both nodes of a pair of clustered VMs, and if the Hyper-V host fails, the SQL will continue operating on the surviving VM. But there may be a very brief period of unavailability while the surviving SQL VM takes ownership and starts serving out requests. Neither the OS or SQL would have to restart in this environment, but it does take just a bit of time to transfer ownership of the resources.
Q: Are there any fault-tolerant solutions for Hyper-V?
| Windows IT Pro Aug. 4, 2012
A: Fault tolerance allows a virtual machine (VM) to carry on running without interruption, even in unplanned host failures such as a host crashing. This is different from high availability, part of Hyper-V, which in a host failure moves VMs to another host but has to restart the VMs, incurring a small outage to the VM. This is also different from planned outages, which allow a VM to be moved between hosts with no downtime using technologies such as Live Migration.
This fault tolerance is achieved by the VM running on multiple hosts with changes from the master replicated in real time to the slave. It should be noted that fault tolerance protects only from a crash of the host; any problem within the guest OS isn’t protected by fault tolerance as any guest problem would just replicate to the copy.
Hyper-V doesn’t have a built-in fault tolerant solution, but there are some options from third parties you can evaluate. However, typically fault tolerance of an application is better handled through application-aware solutions or guest clustering, that provide protection from guest OS crashes. (A good discussion of this can be found at this MSDN blog.) The two main third-party solutions are as follows:
Stratus Launches ’Mission-Critical Hyper-V’ across …
MAYNARD, Mass., USA and LONDON, UK, Oct. 6, 2010 -Expanding the options for businesses seeking affordable uptime reliability for demanding virtual workloads, Stratus Technologies today announced support for Microsoft Hyper-V across its entire ftServer line of fault-tolerant platforms. Mission
. . .
the company went a step further by announcing support for Hyper-V. This means that the Microsoft hypervisor gets out-of-the-box the famous Stratus’ 99.999% uptime.
Specifically, Stratus now supports Windows Server 2008 R2 on its entry level fault tolerant system ftServer 2600.
. . .