• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

anyone managing hyper-v clusters?

yup - that's my primary job.... What kind of info? memory over committing? Dynamic vhd's?

Clustering with Hyper-V can be "finicky" I guess in regards to SAN types - ATM you'll get best bang for buck on 8gb fiber SANs or better (ease of setup, live migration times, amount of maintenance, etc etc), although it does work OK with iSCSI.

I started working with VM's in ESX, then VSphere and frankly, Hyper-V has become a must with all things taken into consideration including licensing.

With the massive core changes to SCOM, VMM and Server8 due out this summer, IMHO M$ is going to become the de facto VM host platform for both VServers and hosted desktop pools....it's that good.
 
I manage both vSphere and Hyper-V. Microsoft needs to seriously bring it to compare with vSphere, they need to step up the game considerably to compete with the feature set ASIDE from licensing. Unless they bring a hypervisor with little overhead, meaning dropping dom0, then it's going to take a whole lot to compete.

That's not to say Hyper-V doesn't have it's place. I think it's a wonderful environment for Dev, that's what we run. I have two clusters running over 200 three-tier VM's and it works decently enough managed via SCVMM w/Self Service Provisioning for the Dev team. Now..would I run 200 Prod Server VM's on Hyper-V in it's current state and comparing to the feature set of my vSphere environment, that's another story.
 
I have ~10 servers. 2 low impact SQL Servers and 2 app servers that require 99.9% uptime.

they are currently running on 2-3 year old PE710's, and none of them get even close to taxing the machine. the busier SQL server, I did some perfmon collection, and in about 2 weeks it only taxes more than 1 core past 50% twice, and then it was only 3 cores over 50%(all 3 are dual quad cores) and none of them seem to use more than 7gb of ram, ever.

the web, file server, fax server, and the few others are not mission critical.

I have about 60-70K in budget for servers coming up, and am starting to try and figure out if I could get maybe, 4-5 dual quad core xeon 710's with 48gb of ram or so and go virtual. but Storage costs and licensing(if going VMWare) could make it not doable.

sadly of the 4 710's I have, I have 2 differenet matched pairs. and all are old enough I cant just buy a few more

and I really know very little about big network disk arrays. I would need to have purchasing sorted out by.......april proably as I need to pull the trigger by June, and would have to do a RFP/Bid thats open for 30 days(government, I cant be trusted to make the correct purchase)
 
Psykos0unds said:
yup - that's my primary job.... What kind of info? memory over committing? Dynamic vhd's?

Clustering with Hyper-V can be "finicky" I guess in regards to SAN types - ATM you'll get best bang for buck on 8gb fiber SANs or better (ease of setup, live migration times, amount of maintenance, etc etc), although it does work OK with iSCSI.

I started working with VM's in ESX, then VSphere and frankly, Hyper-V has become a must with all things taken into consideration including licensing.

With the massive core changes to SCOM, VMM and Server8 due out this summer, IMHO M$ is going to become the de facto VM host platform for both VServers and hosted desktop pools....it's that good.
Click to expand...

does failover work well? automatically?

Vader said:
I manage both vSphere and Hyper-V. Microsoft needs to seriously bring it to compare with vSphere, they need to step up the game considerably to compete with the feature set ASIDE from licensing. Unless they bring a hypervisor with little overhead, meaning dropping dom0, then it's going to take a whole lot to compete.

That's not to say Hyper-V doesn't have it's place. I think it's a wonderful environment for Dev, that's what we run. I have two clusters running over 200 three-tier VM's and it works decently enough managed via SCVMM w/Self Service Provisioning for the Dev team. Now..would I run 200 Prod Server VM's on Hyper-V in it's current state and comparing to the feature set of my vSphere environment, that's another story.
Click to expand...


does failover work well? automatically?

does it have a feature like vmotiion? not that it would be necessary, just 'nice'

can I set priority on VM's so my main AP and 2 sql servers are always up unless shit really hits the fan?

I do not need that much storage. I currently back up pretty much everything to a 2tb raid array and have room to spare.
 
HA works fine and it does have a feature like vmotion, it's called Live Migration. There isn't a feature like Fault Tolerance with Hyper-V where you have "mirrored" system standing by to provide 100% uptime. You can do "guest clustering" with your SQL guests, but currently you are bound to using iSCSI. Windows Server "8" is supposed to provide passthrough HBA support.
 
Vader said:
HA works fine and it does have a feature like vmotion, it's called Live Migration. There isn't a feature like Fault Tolerance with Hyper-V where you have "mirrored" system standing by to provide 100% uptime. You can do "guest clustering" with your SQL guests, but currently you are bound to using iSCSI. Windows Server "8" is supposed to provide passthrough HBA support.
Click to expand...

so do physical clustered machines not share load? or am I misunderstanding?
 
Microsoft uses Failover Clustering and it's only active/active when you run different SQL instances. For instance, if you have a two node SQL Cluster, you can run SQL in the MS Failover Cluster, but you can run one instance on node one and another on node two, and have each node be failover to each other but only for those specific instances. If you want SQL with optimal uptime you can run SQL mirroring where you have duplicate data on another SQL box.

This is how you are spreading the load for the SQL application but not the instances..does that make sense?
 
yeah that makes sense.

its definately a + for VMWare to just have that machine move on over and nevereven appear to be down, but it has a nice price tag.

the software can run for like 2 weeks without the server active but you cant make any changes, you are stuck in the state that the server went down in.

we have a microwave array to another site, but its currently capped at like 4 meg, and if we can come up with like 80K we can upgrade to 10. in the tower base station across the parking lot I have 100meg fiber, thats 2 cards away from being gig. not the best backup place since the network all is in teh same room but atleast teh data is safe if the buidling burns down

these SQL instances do not push tons of data, so it might be doable over the microwave

in the next 1-2 years i will have fiber lit to that site though
 
You must log in or register to reply here.
Back
Top