3

Suppose I have a blade server HP C7000, with three blades.

Q1. Is there any disaster recovery technique that provides redundancy between blades? If any one blade goes down the second should come up with the same configuration.

Q2. Is it possible to enable Memory and storage sharing between blades? If I observe higher usage on a blade and I want to use second card's memory as extending the first blade's resources.

Q3. Redundancy in virtualization - Which applications and technology can be used with redundancy?

Gaurav Arora
  • 31
  • 2

3 Answers3

5

Is their any disaster recovery technique to make redundancy between blades?

Replace "Blade" with "Server" and you get the answer. Blades are independent servers. Standard technologies do not become invalid just because you pack them dense into a blade center.

Can it be possible- Memory and storage sharing between blades?

Yes, it is possible - shared memory between computers as single instance image exists. It is not feasible, though, and a nieche technology, mostly limited to special very rare Hypervisor-Scenarios. Generally "plan your servers as you need them" is the solution.

If you want to research such technology, start at http://en.wikipedia.org/wiki/Single_system_image - there is also a list of software as a starting point. Again - this normally makes little sense. Do not even start thinking unless you have a Infiniband QDR backplane.

Redundancy in virtualization- Which applications and technology can be used with redundancy?

The same as with any other server. Again, blades are not anything else than a different physical packaging of servers. Use the normal technologies that you know as a professional admin and just ignore that they come in a special form factor.

THe CHassis itself will rarely fail - I use a Dell 1000e myself ans seriously, all active components are redundant (6 power supplies, 2 chassis controllers, 9 Fans). What can fail is the KVM master module and the backplane. I dont care about the KVM (does not stop blades from working) and the backplane has no active elements.... so it is unlikely to just fail.

TomTom
  • 50,857
  • 7
  • 52
  • 134
1

  • Q1: This depends on what you have running atop the blade servers. If it's a virtualization suite (VMware, Hyper-V, KVM), then you have some clustering and high-availability options. Same for databases with clustering capabilities.

  • Q2: No. Not really for your use case.

  • Q3: See Probability of Blade Chassis Failure

Community
  • 1
ewwhite
  • 194,921
  • 91
  • 434
  • 799
0

Q1. Is their any disaster recovery technique to make redundancy between blades? If any blade goes down and second should come up with the same configuration.

Not that are blade specific no, but blades are great, especially if you buy a lot of servers - we've been buying C7000's since the day they were announced and love them to bits - but they're kind of dumb in a way, which actually HELPS with resilience rather than hinders.

Q2. Can it be possible- Memory and storage sharing between blades? If I observe higher usage on a blade and I want to use second card's memory as extending the first blade's resources.

Again not as part of the blade environment no, nor would you want them to - but they support any other high-availability functionality any other setup supports.

Q3. Redundancy in virtualization- Which applications and technology can be used with redundancy?

There are lots of resilience functions built into any hypervisor - we can't go into it here but we use VMWare vSphere/ESXi on almost all of our blades and their HA/DRS/FT/sDRS etc. functions just perfectly on our blades.

Basically don't sweat it - they're great, there's lots of us on this site that can help you too if you go ahead with blades - they're just a form factor, nothing to worry about.

Chopper3
  • 100,240
  • 9
  • 106
  • 238