Disaster Recovery - Hosted Server
-
@scottalanmiller It's the cost of new server hardware + the 4 hour warranty callout each year.
Plus if a raid controller breaks, you still have to wait for the hardware warranty + rebuild the array and then restore the data.
-
@thecreativeone91 Cisco...Dear god no.
Dell? Yes I've seen their warranty guys in action but still, for a service as simple as active directory and DNS, can we still not use hosted services as a redundant backup?
-
Also, let's assume some sites are happy to pay for COLO and other such methods if needs be.
Having a redundant server in the same room is not a great idea, power cuts, floods, ect.
That 4 hour SLA from Dell is worth nothing if power to the street is down for 2 days.
-
@Breffni-Potter said:
@scottalanmiller It's the cost of new server hardware + the 4 hour warranty callout each year.
Plus if a raid controller breaks, you still have to wait for the hardware warranty + rebuild the array and then restore the data.
Nope, not if the RAID controller goes. Just need to swap the part. We could do that in under fifteen minutes from call to replacement when we had HP! FIFTEEN MINUTES! That's pretty crazy but they kept parts stocked nearby, ran them over on foot and replacement was minutes. Few things recovery as quickly as a blown controller because if that is all that dies, it's literally just power down, plug in, power up. Done.
-
@Breffni-Potter said:
@thecreativeone91 Cisco...Dear god no.
Dell? Yes I've seen their warranty guys in action but still, for a service as simple as active directory and DNS, can we still not use hosted services as a redundant backup?
If it is JUST the services, why not run them in the cloud full time? That's what we do. We have AD and DNS out on Azure. So we can fail with application level failover and no one would notice.
-
@scottalanmiller I don't disagree, this is what I am trying to look into, I don't trust Azure 100% yet so wondering about whether a hybrid approach would work.
On site DC with hosted offsite DC.
-
@Breffni-Potter said:
@scottalanmiller I don't disagree, this is what I am trying to look into, I don't trust Azure 100% yet so wondering about whether a hybrid approach would work.
On site DC with hosted offsite DC.
Of course, either works. Although I'd have the opposite reaction - I don't trust your site, Azure I trust
Having on site and Azure DCs in an application cluster is completely transparent. No way to know when you are using which except that there would be higher latency on the connection to Azure.
-
We used to have AD and DNS on site. Now we have then in Azure East (VA) and Azure Central (Iowa).
-
@thecreativeone91 said:
Easy, Do 4 hr or less on site service contract (both dell and cisco do 2 & 4 hr).
The have baremetal backups.This only promised that they will be at your door in 2 or 4 hours.. not fixed in that timeframe!
-
@scottalanmiller How do the Azure VMs connect to each site? VPN to each site? or does each site VPN into the Azure instance or what?
-
@scottalanmiller said:
@Breffni-Potter said:
@thecreativeone91 Cisco...Dear god no.
Dell? Yes I've seen their warranty guys in action but still, for a service as simple as active directory and DNS, can we still not use hosted services as a redundant backup?
If it is JUST the services, why not run them in the cloud full time? That's what we do. We have AD and DNS out on Azure. So we can fail with application level failover and no one would notice.
how does licensing working for that?
-
I imagine you use Pertino to do the connection between the different clients with ADconnect?
What if you were to do it without Pertino?@Dashrender said:
how does licensing working for that?
As far as I can tell, Azure says you do not need to buy CALS, you pay for it with their hosting package. So the per hour fee you pay covers the server license along with hardware/bandwidth.
-
What if you could keep the Azure VM off as a disaster box, power it on, restore from backup, would be much cheaper long term if you only paid for Azure storage rather than have it on 24/7 then if a disaster does strike you can use it.
-
@Dashrender said:
@thecreativeone91 said:
Easy, Do 4 hr or less on site service contract (both dell and cisco do 2 & 4 hr).
The have baremetal backups.This only promised that they will be at your door in 2 or 4 hours.. not fixed in that timeframe!
That's what the six hour "fix" agreement is for. That's the best one.
-
@dafyre said:
@scottalanmiller How do the Azure VMs connect to each site? VPN to each site? or does each site VPN into the Azure instance or what?
VPN. For us, it's Pertino. But you can use anything that you want.
-
@Dashrender said:
@scottalanmiller said:
@Breffni-Potter said:
@thecreativeone91 Cisco...Dear god no.
Dell? Yes I've seen their warranty guys in action but still, for a service as simple as active directory and DNS, can we still not use hosted services as a redundant backup?
If it is JUST the services, why not run them in the cloud full time? That's what we do. We have AD and DNS out on Azure. So we can fail with application level failover and no one would notice.
how does licensing working for that?
Same as anything else. Server license is provided by the hosting agreement. You provide your CALs.
-
@Breffni-Potter said:
As far as I can tell, Azure says you do not need to buy CALS, you pay for it with their hosting package. So the per hour fee you pay covers the server license along with hardware/bandwidth.
Is that true? Unlimited CALs as long as you use all Azure? What about if you have some on site machines? What if you are a 400K person company, have millions of dollars per year in CAL costs, but have ONE Azure instance, just big enough to fire up Windows. Are all of your CALs suddenly free?
Doesn't seem possible.
-
@Dashrender said:
@thecreativeone91 said:
Easy, Do 4 hr or less on site service contract (both dell and cisco do 2 & 4 hr).
The have baremetal backups.This only promised that they will be at your door in 2 or 4 hours.. not fixed in that timeframe!
True. Especially with Dell. With Cisco SmartNet OnSite if you have either servers or networking devices you can have them backed up to the cloud (as well as local) and Cisco will just have someone their with a replacement unit in 2hrs and will reload the config from backups.
-
@Breffni-Potter said:
What if you could keep the Azure VM off as a disaster box, power it on, restore from backup, would be much cheaper long term if you only paid for Azure storage rather than have it on 24/7 then if a disaster does strike you can use it.
Can you? Yes. How much will you save? Some. Is it worth it? I doubt it. I'd put this as getting pretty close to silly in terms of cost savings versus value. It this was a different workload, there could be benefits. But we are talking about AD that works just as well from Azure as on site and works BY FAR The best when kept in sync. If you keep the VM running, you get transparent failover 100% of the time. And you get to test it every day and know that things work.
-
@Breffni-Potter said:
As far as I can tell, Azure says you do not need to buy CALS, you pay for it with their hosting package. So the per hour fee you pay covers the server license along with hardware/bandwidth.
I've never seen that before. You need CALS not matter where the server is located.
