Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

I was wondering, how does a 'traditional' virtual machine setup with storage on iSCSI, with a fallback using DRBD, nodes with Xen installed the 'normal way', storing their images on the iSCSI target, etc. compare to Cloud solutions, like Ubuntu Enterprise Cloud?

Researching about clouds doesn't seem to give me much concrete information. I still have a bunch of uncertainties about is. For example, does Ubuntu enterprise cloud have integrated redundant/fall back storage and failover VM nodes? Can storage be expanded as simply as adding another storage node?

Let me ask it this way: is it unwise to implement a xen setup using the aforementioned tools now that Cloud stuff is gaining popularity? Will the 'old' way be deprecated soon?

share|improve this question

2 Answers 2

Are you asking if you could setup a fault tolerant cluster locally with open source tools cost-effectively, compared to purchasing cloud instances from a 3rd party?

I think it depends on how good you are :-)

A major benefit of outsourcing is that they specialize in the hardware plumbing and fault tolerance, so they could be much more efficient at it.

One day everything may be in the "cloud". There are a few things stopping me from moving the majority doing so yet:

  • Our corporate email server has gone down less often then my personal gmail/gcal account.
  • Having servers locally gives insight into failures; appose to occasional updates from a provider during an outage. Even if it only gives a sense of control (ie, you couldn't fix it faster), that sense can be important.
  • I've done the math for a few of our services; we still come out cheaper
  • Many services are still valuable during an internet outage, if you have local users.

That said, as we're rolling out new services, I'm evaluating 3rd parties more and more regularly.

These items may or may not be applicable to you, its really dependent on the application.

If you're asking about using UEC locally versus Zen; local UEC is just a system built around KVM virtualization. I can't say which is better. You could ask "I need feature X,Y, and Z; should I use UEC or roll my own zen?"

share|improve this answer
    
I know there are 3rd parties which have excellent setups, but I'm just asking about setting up one myself. My question mainly boils down to how robust UEC is. My experience isn't too good, as the web interface borked out on me with a fatal error in a clean install when I tried it first... Anyway, I can't really ask "I need feature x, what do I use?" I was mainly wondering about its fallback/failover features. When a node fails, is there fallback? Is it easily repaired, or will it be like a Windows machine which is impossible to fix with anything but a reinstall? –  Halfgaar Feb 16 '11 at 11:26
    
Perhaps you should ask a new question "Is UEC production ready?" Then add "How does it handle node failure" and mention your error during the clean install. I think you'll get some better answers that way. –  Steven Feb 16 '11 at 18:42

Agree with all that was previously stated. In addition you should look closely at FC SAN storage versus iSCSI. As a Data Center, our customers have tried several iSCSI systems. Most have had real problems. Corrupting databases, incompatibility with switches, Ethernet Bonding, and driver issues. The FC SAN systems have all worked with little to no problems.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.