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

We recently had a disastrous transition from a single server (Win2008 server, 4 dual core 2.8GHz processors, 24G RAM) to a much larger, faster Hyper-V server. We run Apache 2 and PHP 5 (both 32-bit) and really nothing else.

The new VM has a guest with a Win2008 server operating system with an identical configuration as the original server. The entire guest OS and Apache server root is stored on a SAN with a 2G NIC (1G read and 1G write). I am told that is required in order to have the ability to failover to another VM if needed.

When running Apache Bench with identical parameters, the old single server performs 20 to 30% better than the new "bigger, faster" Hyper-V machine. We installed a Linux box with Apache 64bit on the VM and still had the same performance issues.

I can't figure out why a machine that is in all areas a "much better" machine (RAM, processors, etc.) would perform so poorly. Is this a result of having the operating system and Apache root stored on the SAN? Does a Hyper-V installation always perform worse than a non-virtual server? Where would you look and how would you determine what the problem is?

share|improve this question

1 Answer 1

up vote 2 down vote accepted

The new VM has a guest with a Win2008 server operating system with an identical configuration as the original server.

How did you do that given that Hyper-V only supports 4 virtual processors before the version currently in beta, so you are limted to 4 cores on the VM?

I am told that is required in order to have the ability to failover to another VM if needed.

No, you are told it is a requirement to fail over to another HOST - not another VM. PLEASE try to listen to what people tell you.

Is this a result of having the operating system and Apache root stored on the SAN?

You tell us. Theoretically this is 100mbit per second - plenty, But who knows how fast the SAN is? Run low level comparisons and find out whether the SAN performance is much worse than the performance of the old disc subsystem. Not a joke - seen that.

Also note that you have a lot less CPU cores available.

Does a Hyper-V installation always perform worse than a non-virtual server?

Depending on load - yes. See, multi VM's means less second and third level cache hits (as the CPU jumps between programs). Not Hyper-V relevant - this is typical for diverse loads. But it should not be so obvious, especially if you are the only machine on the host. You really need to sit down and do basics analysis for sys admins. Find out how loaded both the host and the VM are, run comparisons.

In general it should not be the case -especially given the pathetic hardware you started with (4 dual core old), but you DID lose CPU cores on the transition.

share|improve this answer
    
I'll accept that answer, though don't know why you had to throw in the sarcasm. You don't know what "people tell" me. If they had told me correctly and knew the answers, I wouldn't have been asking on this forum. I understand that a lot of idiots post here, but don't just make the assumption that this poster is one of them. Thanks for the answer though. –  Jonathon May 2 '12 at 19:10

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.