From the Chief Technology Officer at Marathon Technologies

Jerry Melnick

Subscribe to Jerry Melnick: eMailAlertsEmail Alerts
Get Jerry Melnick via: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn

Related Topics: VMware Journal, Citrix Virtualization Journal, Infrastructure 2.0 Journal, Mixed Network Integration, Microsoft Developer

Windows 2008: Blog Feed Post

Q & A for the January 2009 Webinar—Customer Spotlight: How the Sullivan Group Got Reliable High Availability without Breaking the Bank

We had a lot of great questions during the Q & A session of our January webinar with one of our customers. We’ve posted the questions and responses here on our blog for everyone’s benefit.

Questions for The Sullivan Group:

Q: Which everRun product are they running? everRun HA or FT?
A: everRun VM

Q: How did you migrate your VMware VMs to XenServer?
A: We used a V2V software migration tool from Visioncore that worked really well.

Q: What was the procedure when bringing back up one of the servers when the RAID card failed. How easy was it?
A: It couldn’t have been easier. Once we repaired the failed component, everRun identified it and put it back into use. everRun used a mirror copy to bring the two systems back in complete synchronization. This all happened without our intervention and without impacting our users.

Q: Was a short implementation time a demand from The Sullivan Group, and what was the expectations before the implementation?
A: Short implementation time wasn’t a hard requirement. We expected the implementation to take a couple of weeks and were pleasantly surprised when we had it all up and running in a little less than a week.

Questions for Marathon:

Q: Is there a special license for SQL? Or any other special considerations?
A: Special licensing for the applications you protect is not required. You should refer to the SQL license agreement or the agreement for the application on the VM that you are protecting. With everRun, only one instance of the application is running at any one time.

Q: Is this active active or active passive?
A: We refer to it as ‘active/ready’. The secondary VM is in a paused state, however disk and network I/O are being processed. This allows everRun to deliver fault tolerance at the component level, while immediately starting the paused VM fully in the event of a full host failure on the primary side.

Q: Do you have to install everRun prior to your application, so if I already have an application installed would I need to rebuild the production server? IF so does this change the install time and impact?
A: Assuming you have a VM set up on a host, the application can be installed before or after the VM is protected with everRun. The protection process does require the VM to be shutdown prior to beginning the protect process, however the process takes about 2 minutes to complete after which time the VM can be restarted.

Q: How does everRun handle Software or OS hangs?
A: everRun does not monitor applications.

Q: In your opinion what is the strongest difference between this and MS Clustering?
A: MS Clustering can be a nice fit, especially with applications such as SQL Server. One of the requirements for MSCS is shared storage, or a SAN. This requirement can push the cost out of reach for many small and midsized businesses. everRun does not require shared storage and can utilize any type of storage the customer may have or intend to purchase. In addition, everRun provides fault tolerance and not just failover restart. This helps to minimize interruptions typically caused by failed devices.

Q: Does everRun VM support Windows x64 architecture on VM?
A: Yes, everRun and XenServer support 64-bit hardware and software.

Q: Can VM run Windows 2008 x64? Or Windows 2003 x64?
A: Citrix XenServer supports VMs running Windows 2008 64-bit and Windows 2003 64-bit. Currently everRun VM can protect VM’s running Windows Server 2003, 32-bit and 64-bit. Our next release planned for calendar Q2 09 will support Windows Server 2008 32-bit and 64-bit VM’s.

Q: How much overhead do you have when you protect a VM?
A: When protecting a VM you are able to define how much of a particular resource is to be utilized and reserved. This helps to reduce the amount of resources required for protecting VM’s. Performance overhead can vary depending on if it is I/O heavy, CPU heavy, what the application is, etc. Typically however performance overhead is not impactful.

Q: Is everRun tied to a VM on a particular physical XenServer, what would happen if you used XenServer technology to move to another physical server?
A: Once a VM is protected with everRun, it is hard-configured to two physical hosts. everRun allows online migration of the active VM between these two hosts without interruption. To move one or both VM’s in the protected pair to a different host would require it to be unprotected, migrated if necessary, and protected again with the new host.

Q: Please speak about XenServer integration, process for failover to DR site?
A: everRun has a tight integration with XenServer. everRun is installed on top of XenServer and is completely compatible and able to protect Windows VMs created in XenCenter. During the protection process, everRun takes the chosen VM and clones it to the designate secondary host. This creates a complete and identical VM on the secondary host. everRun maintains these two VM synchronously so that they are always identical. everRun’s unique architecture exposes these two mirrored VM’s as a single VM; there is no need to install, manage, or update both sides, only the one single instance of the OS/application. Should the entire ‘primary’ host fail, the ‘secondary’ host will immediately start the cloned VM. It comes up with the same IP address, hostname, and MAC address of the primary so that there are no client-side, DNS, Active Directory, or other infrastructure changes required.

Q: Will everRun run on x64?
A: Yes, it requires 64-bit servers.

Q: I assume the servers must match i.e. memory, HD space and memory as well as other array controllers and type of arrays?
A: The only requirement for similarities within the servers is same family of processors. everRun can mirror storage between dissimilar storage types and vendors, allowing lower-cost storage to be deployed on the ‘secondary’ host.

Q: Is everRun for virtual servers XenServer specific, or does it work with Hyper-V, Virtual Iron, etc.?
A: Today everRun is developed for XenServer. In January we announced a development and marketing agreement with Microsoft, we will be developing an everRun product for Hyper-V as well.

Q: What if the server is up but a single app on the server fails? Also, how do you detect the app failed?
A: everRun does not monitor applications. However if by ‘single app’ you mean a single VM, everRun does protect at the individual VM level. If a VM fails yet the host and other VM’s remain alive, everRun can restart the VM on the secondary host.

Q: Do you have to purchase redundant licenses for the applications that you have replication as Virtual Machines across two physical servers? For example, do you have to purchase double the Exchange licenses to do it or just the licenses as if you had a single server?
A: You would need to refer to the license agreement for each application. With everRun, the application is running as a single instance and many vendors don’t require two licenses, but this varies between vendors.

Q: Can you have full fault tolerance on XenApp servers between 2 datacenters, where users would not be disconnected from their session if a failover occurred?
A: everRun can certainly protect XenApp as it can protect any Windows application. A number of customers are using everRun to protect XenApp today. The ability to separate between data centers (there are latency requirements due the synchronous nature of everRun) will be available in Q2. To prevent session disconnects will require Level 3 protection, or full System-Level Fault Tolerance, which will also be available for everRun VM in Q2.

Q: Can the VM servers be in different data centers across a WAN behind firewalls?
A: The ability to separate servers geographically will be available in Q2.

Q: Are there general guidelines on the number of VM's that can be protected between two host machines? I'm thinking of SQL Server systems hosting highly transactional databases.
A: The answer to this is dependent on the applications running within the VM’s, hardware, and activity within them, so providing a set number of VM’s would not be practical. Please contact Marathon if you would like to discuss your environment to better understand what VM limits may be suitable.

Q: How can you use the USB interface, for example when software requires a dongle?
A: everRun does not redirect the USB interface to a protected VM.

Read the original blog entry...

More Stories By Jerry Melnick

Jerry Melnick (jmelnick@us.sios.com) is responsible for defining corporate strategy and operations at SIOS Technology Corp. (www.us.sios.com), maker of SIOS SAN and #SANLess cluster software (www.clustersyourway.com). He more than 25 years of experience in the enterprise and high availability software industries. He holds a Bachelor of Science degree from Beloit College with graduate work in Computer Engineering and Computer Science at Boston University.