Go Premium for a chance to win a PS4. Enter to Win

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 347
  • Last Modified:

high availability of physical servers

Is there anyway to build in fault tolerance/high availability into physical servers? Our IT section has a core applicaiton server that is still a physical - for virtual machines if the host they were on blew up - the HA cluster would restart them on another host, but for physical servers - what kind of HA safeguards can you build in if any? Or are they single points of failure?
0
pma111
Asked:
pma111
  • 11
  • 6
  • 2
  • +2
3 Solutions
 
strivoliCommented:
Cluster is the solution and depends on many factors. First of all the OSes involved. Basically a Cluster is built from 2 servers that supply the same service/application. When one fails the other runs. There are active/active and active/passive Clusters.
0
 
pma111Author Commented:
But no cluster = no safeguards if the physical server dies?
0
 
pma111Author Commented:
Can you elaborate on this:

>.First of all the OSes involved.

>.There are active/active and active/passive Clusters.
0
Get your Disaster Recovery as a Service basics

Disaster Recovery as a Service is one go-to solution that revolutionizes DR planning. Implementing DRaaS could be an efficient process, easily accessible to non-DR experts. Learn about monitoring, testing, executing failovers and failbacks to ensure a "healthy" DR environment.

 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
If you use Microsoft Failover Clustering, there is a short delay, while the process is transferred to the second server in the cluster.

I would recommend you take a look at a product called Double-Take HA

Real-time High Availability: Total workload mobility and protection for unlimited IT evolution

Double-Take® Availability for Windows 7.0 is the market’s first real-time high availability and disaster recovery solution to offer agentless and agent-based protection options for physical, virtual or cloud servers. Fine-tune your recovery and efficiency targets with flexible RTO and RPO.

No matter the size of your business, now you can have full-scale protection at one affordable price. Double-Take Availability is hardware agnostic and reduces your overall operating costs by optimizing existing infrastructure and scaling with whatever you build in the future.

Source
http://www.visionsolutions.com/Products/DT-Avail.aspx
0
 
pma111Author Commented:
What are the general ballpark costs for a high spec applicaiton server (physical) in todays market?
0
 
pma111Author Commented:
Wondering if the lack of cluster setup may be due to limited resources £/$
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
Availability is just about, money, how much downtime can the business afford, and the technical ability to implement a solution. (which could be money, if your need Consultancy!)

Spend more money on a solution, and obtain high availability.

e.g. simple downtime, you will need an additional server, and additional Microsoft OS, for Clustering, with 1-2mins of downtime, zero downtime, more money.
0
 
pma111Author Commented:
What are the general ballpark costs for a high spec applicaiton server (physical) in todays market?
0
 
pma111Author Commented:
And if you have no cluster setup - can you give an expert opinion how long it takes to rebuild a physical server from scratch, i.e what considerations are there on this front as to how quickly you can rebuild and get back up and running, are you talking days, weeks, longer?
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
At a rush, it would take us 5 hours, that's from inserting DVD/CDROM, Applying the same Security Patches, based on Windows 2008 R2, and applying specification configuration.

That is just a base OS, install 4 hours+patches, 1 hour configuration.

That's 5 hours to complete, not 5 man hours of effort.

Considerations, what does the server do?
0
 
pma111Author Commented:
Do you not have to wait for the new kit to arrive though? I.e. if a physical server dies, is there not also the period of time where the new hardware arrives - what are common turnaround times for that?
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
That's assuming we have hardware in stock, which we do! (in the warehouse!), of all production systems for clients! (New, Sealed in Boxes!).

Lead times for hardware replacement, and assuming different hardware, that 5 hour, could be weeks.

Procurement time - 2 weeks, delivery of hardware is not uncommon, for Dell or HP, IBM.

Of course, most hardware is under Support and Service Agreement with Dell/HP, on a 1 or 4 hour turnaround.
0
 
strivoliCommented:
If you don't have a HA solution, you should at least set up a Disaster Recovery (DR) policy. The DR documents what steps must be performed in order to make the Server UP again. Furthermore, the DR also estimates how long the Server will be DOWN in case of different failures.
0
 
pma111Author Commented:
>Of course, most hardware is under Support and Service Agreement with Dell/HP, on a 1 or 4 hour turnaround.

What does this mean, i.e. 1 hour to get the actual hardware to your office? or what is the 1-4 hour turnaround specific too?
0
 
Mohammed KhawajaCommented:
There are three options which could deliver what you are looking for:

1.  Software solution such as DoubleTake or in some circumstances, clustering
2.  Lockstep technology hardware from either NEC, Bull or Stratus
3.  Stratus Avance software

Lockstep technology in essence is two fully redundant servers where both are executing the same instructions.  When one fails, the other takes over (usually in 5 seconds or so).

If the application/system is so critical then what are your plans if you are using shared storage (i.e. SAN failure).  Stratus deals with this too as they have storage arrays (SAN) which are fully redundant just like their servers.  It is pricey but it is bullet proof.

Stratus Avance is new and it works but I have not had the opportunity to test it.

Check out more High Availability options from Stratus by going to http://www.stratus.com
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
Log a call with HP, and they will supply and engineer and parts to fix your server within 1 or 4 hours, depending upon Support Contract.
0
 
andyalderSaggar makers bottom knockerCommented:
What OS/application is involved, there may be a cluster solution at the application level, e.g. Oracle RAC.
0
 
pma111Author Commented:
OS is WIndows Server but it isnt a well known application.
0
 
pma111Author Commented:
I'm confused, this hints at a new server would take 2 weeks to arrive, but then the bottom bit says it would only take 4 hours to get the hardware. Or are you saying once purchased (i.e. the 2 weeks), then it is only 4 hours for the admin to arrive with it and fit it for you. So in total 2 weeks and 4 hours.


Procurement time - 2 weeks, delivery of hardware is not uncommon, for Dell or HP, IBM.

Of course, most hardware is under Support and Service Agreement with Dell/HP, on a 1 or 4 hour turnaround.
0
 
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
You either

1. purchase a new server if you have NO SUPPORT CONTRACT

2. Get it repaired if you have a SUPPORT CONTRACT
0
 
pma111Author Commented:
understood, many thanks
0

Featured Post

Automating Your MSP Business

The road to profitability.
Delivering superior services is key to ensuring customer satisfaction and the consequent long-term relationships that enable MSPs to lock in predictable, recurring revenue. What's the best way to deliver superior service? One word: automation.

  • 11
  • 6
  • 2
  • +2
Tackle projects and never again get stuck behind a technical roadblock.
Join Now