[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-users] HA problem

  • To: <Xen-users@xxxxxxxxxxxxxxxxxxx>
  • From: "Gustavo Pimentel" <gustavo.pimentel@xxxxxxxxxx>
  • Date: Thu, 9 Jun 2011 17:30:48 +0100
  • Cc:
  • Delivery-date: Thu, 09 Jun 2011 09:32:33 -0700
  • List-id: Xen user discussion <xen-users.lists.xensource.com>
  • Thread-index: AcwmwpaHYmkBOwVNQc6PDJgSkmFeLw==
  • Thread-topic: HA problem

Hi guys,


I need you help for a HA problem. I have to implement a equipment with HA solution based on 2 servers, I’m sure you could help.


The basic idea is something like this:

·         On both servers I was thinking to install Debian 6.0.1 “Squeeze” with xen 4.1.0 hypervisor and xen-tools.

·         Creating a virtual machine on Server 1 and replicated to Server 2 only for the first time.

·         Running the virtual machine on Server 1 and try through remus and DRDB send checkpoints and disk modifications to Server 2 and hope a virtual machine be created on Server 2 on hot standby role.

Case 1.png

·         Using HeartBeat on both servers I could detect a failure on each server.

·         If the Virtual Machine on Server 1 or the Server 1 itself suffer some kind of failure, I would expected the virtual machine on Server 2 change from hot standby to online role.

Case 2.png

·         Meanwhile if the Server 1 recovers from the failure, it should now receive the checkpoints and disk modification from Server 2 becoming this assuming the hot standby role.


My questions:

·         It’s possible to do this with remus and DRDB? On http://nss.cs.ubc.ca/remus/doc.html webpage it seems so. But can anyone confirm that?

·         For now this test will be performed on a computer with CPU Intel Pentium 4 at 3.2GHz (32bits) without KVM support and with 2GB of RAM. Does anyone knows if this type of machine could has problems with xen?




Xen-users mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.