?
Solved

Hyper V Live migration with limited Memory

Posted on 2011-10-29
4
Medium Priority
?
288 Views
Last Modified: 2012-05-12
Hi,
I have 2 Hosts install Hyper V 2008 R2 and clustered.
both Host with 12G memory.
if I have 2 VMS running on each Host , and each one configured with dynamic memory 2G-6G.
when they runing normal , each Host have 10G used , each vm used 5G realy memory.
in this case , if one of the host dead , can VMS be live migrated to another Host?
if they all use the minimun , then total 8G memory , it should works , but the real memory used is 10G each and that will not enough .
0
Comment
Question by:GordonLiq
  • 2
  • 2
4 Comments
 
LVL 18

Expert Comment

by:LesterClayton
ID: 37050068
VMs can be live migrated from one host to another provided that there is enough memory available on the 2nd host to accommodate the current memory of the server at failure, plus the buffer (default 20%)

I should point out though that "Live Migration" when a node fails is impossible - because the node is dead.

If the network is lost, the cluster that does not host the quorum disk will be kicked out the cluster.  It has no way to migrate from one host to another (network lost), so the virtual machine is just stopped, and the surviving node has to start it up from scratch, meaning that the startup memory is used.

Similarly, if the disk is lost, the cluster that does not host the quorum disk will be kicked out of the cluster.  It has no way to save the virtual machine state to disk (disk lost), so the virtual machine is just stopped,and the surviving node has to start it up from scratch, meaning that the startup memory is used.

You may end up with an issue if 2 of the virtual machines on the surviving node are both using their maximum amount of memory - then there are no resources for the failed virtual machines to start up.  I'd highly advise to increase the memory of all of your hosts so that it can cope with the maximum amount of allocated memory. You should never over-commit a cluster - always have enough resources available to cope with a complete node failure.
0
 

Author Comment

by:GordonLiq
ID: 37051266
when you said failed, is it different to just restart or shundown?
I have tried with the enough free memory ,   the VM can be live migration (High availability) to another Host, dont know why you said no.

the option of this is I dont want to leave the free memory on each host to be able to HA,   I would like to set the dynamic for each VM (half -full) so that physical memory will always been efficient used.
0
 
LVL 18

Accepted Solution

by:
LesterClayton earned 2000 total points
ID: 37051284
Maybe we need to redefine what "fail" actually means.  In my opinion a failure is when a node dies.  A shutdown or restart of a host is not a failure - it's a controlled method for which a cluster service on a host can automatically migrate running VM's to other hosts.  Technically, the machines do not "fail over" to another node, they are migrated in a safe and reliable way.  There is no HA involved here.  HA is when the guests on the failed host are restarted on the other nodes.

If a Guest has allocated all of the available memory to it (i.e. maximum 6 GB), then it requires 6 GB on the target host to migrate over to.  If it's using the least amount of available memory to it (i.e. 2 GB) then it requires 2 GB on the target host to migrate over to.  

Only you can really determine whether or not the other host can cope with a "failover" be it critical failure, in which case the minimum amount of memory plus 20% is required to restart the guests - or just a restart of the host, in which the currently allocated amount of memory plus 20% is required on the remaining host to migrate the server over to.

If there is insufficient memory available to migrate it, the guest OS is placed in a save state and becomes unusable until the cluster finds enough resources on a host to bring it back up.
0
 

Author Closing Comment

by:GordonLiq
ID: 37051330
If there is insufficient memory available to migrate it, the guest OS is placed in a save state and becomes unusable until the cluster finds enough resources on a host to bring it back up.
this is the answer I looking for
0

Featured Post

Free Tool: ZipGrep

ZipGrep is a utility that can list and search zip (.war, .ear, .jar, etc) archives for text patterns, without the need to extract the archive's contents.

One of a set of tools we're offering as a way to say thank you for being a part of the community.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

Tech spooks aren't just for those who are tech savvy, it also happens to those of us running a business. Check out the top tech spooks for business owners.
Your business may be under attack from a silent enemy that is hard to detect. It works stealthily in the shadows to access and exploit your critical business information, sensitive confidential data and intellectual property, for commercial gain. T…
In this video tutorial I show you the main steps to install and configure  a VMware ESXi6.0 server. The video has my comments as text on the screen and you can pause anytime when needed. Hope this will be helpful. Verify that your hardware and BIO…
If you're a developer or IT admin, you’re probably tasked with managing multiple websites, servers, applications, and levels of security on a daily basis. While this can be extremely time consuming, it can also be frustrating when systems aren't wor…
Suggested Courses

850 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question