What Are You Doing Right Now
-
@art_of_shred said in What Are You Doing Right Now:
@Dashrender said in What Are You Doing Right Now:
@art_of_shred said in What Are You Doing Right Now:
@Dashrender said in What Are You Doing Right Now:
@art_of_shred said in What Are You Doing Right Now:
@DustinB3403 said in What Are You Doing Right Now:
sick day....
trying to sleep more, but head is pounding...
...so you're posting on ML?
If I'm sick enough to be home - you can bet your ass, if I'm awake I'd be on ML.
I'm home every day, but I don't get sick days, so it's all foreign to me. I think I slept in a little once when I was feeling really lousy.
LOL - true. But then again you're on call 24/7. Most SMB IT folks I know aren't.
Not entirely. I make it a point to see everything, 24/7, but only choose to do anything about it M-F/8-5.
What I meant is that you are on call 24/7. Most enterprise IT are on call 24/7, and like you work M-F 7/8-5/6 ish.
-
Right before I left yesterday I overheard a meeting my sort-of supervisors were having in regards to purchasing two new Hyper-V host machines and what their purpose would be. Each server would be identical. One would house our DAG server (it's running out of space anyway) with the other being a replication point for hardware redundancy only, both of which would be Raid 5. This triggered me right before I left for the day.
We are buying two servers and not clustering them so we essentially lose the resources of the second server as it's not being used for anything else. We are also configuring them in a Raid 5 because of budgetary constraints (our client said no so we are trying to find the money in our calculated yearly budget). This is supposed to be a temporary measure until July 1 when we get all of the money for our new budget.
Now, these servers eventually will be repurposed to house about 80% of our infrastructure after July, where the DAG will be moved to a new host that we have yet to purchase. Based on their conversation I don't think they are planning to buy more hard drives in July and configure these hosts in a raid 10 or have they considered clustering at all.
So we're making these things more important with more hard drives and more likely to fail. I need a drink.
-
Assuming they are buying HDDs, you need to show them the math that shows why RAID 5 is dead in that space and that they need to look at RAID 6 or RAID 10.
-
As for storage for your DAG - you need a whole new server for this? what about a used DAS shelf and some disk?
-
@Dashrender The server is on it's way out anyway so it solves two problems but I'm really put out.
-
@wirestyle22 said in What Are You Doing Right Now:
We are buying two servers and not clustering them so we essentially lose the resources of the second server as it's not being used for anything else.
Why would you assume you need/want to cluster them? If you don't have licensing for that, that is yet another expense to make that work.
Also, what Hypervisor are they looking at? Hyper-V would be free in this case, but if they are a VMWare shop, then clustering is even more expense.
-
@Dashrender said in What Are You Doing Right Now:
Assuming they are buying HDDs, you need to show them the math that shows why RAID 5 is dead in that space and that they need to look at RAID 6 or RAID 10.
I really want to avoid parity if this is going to be that big of a percentage of our production servers
-
@wirestyle22 said in What Are You Doing Right Now:
@Dashrender The server is on it's way out anyway so it solves two problems but I'm really put out.
Were you actually in that meeting, or did just you just happen to hear it as you walked by?
-
@dafyre said in What Are You Doing Right Now:
@wirestyle22 said in What Are You Doing Right Now:
@Dashrender The server is on it's way out anyway so it solves two problems but I'm really put out.
Were you actually in that meeting, or did just you just happen to hear it as you walked by?
My desk is directly next to their office so I hear everything
-
@wirestyle22 said in What Are You Doing Right Now:
@dafyre said in What Are You Doing Right Now:
@wirestyle22 said in What Are You Doing Right Now:
@Dashrender The server is on it's way out anyway so it solves two problems but I'm really put out.
Were you actually in that meeting, or did just you just happen to hear it as you walked by?
My desk is directly next to their office so I hear everything
Perhaps you should mention it to them... "By the way, I heard you talking about raid 5..." and then explain to them why it's a bad idea?
-
@wirestyle22 said in What Are You Doing Right Now:
@Dashrender said in What Are You Doing Right Now:
Assuming they are buying HDDs, you need to show them the math that shows why RAID 5 is dead in that space and that they need to look at RAID 6 or RAID 10.
I really want to avoid parity if this is going to be that big of a percentage of our production servers
Really, the thing that matters is IOPs - if the IOP requirement is met by RAID 6, then there shouldn't be an issue. Of course, don't forget to consider IOP in a failed state.
-
@Dashrender said in What Are You Doing Right Now:
@wirestyle22 said in What Are You Doing Right Now:
We are buying two servers and not clustering them so we essentially lose the resources of the second server as it's not being used for anything else.
Why would you assume you need/want to cluster them? If you don't have licensing for that, that is yet another expense to make that work.
Also, what Hypervisor are they looking at? Hyper-V would be free in this case, but if they are a VMWare shop, then clustering is even more expense.
This is a 100% Hyper-V shop.
Isn't one of the strenghts of virtualization that you can allocate resources in a way where wasted resources are at the very least greatly reduced? An almost completely wasted server is the opposite of what we are trying to do in the modern world especially when their reasoning is hardware redundancy. You can have hardware redundancy with clustering and retain all of your resources. Seems like a no brainer to me.
-
@dashrender If they aren't considering all of this I highly doubt they have calculated iops too. I don't get triggered a lot but I feel like we're setting ourselves up to fail.
-
@wirestyle22 said in What Are You Doing Right Now:
@Dashrender said in What Are You Doing Right Now:
@wirestyle22 said in What Are You Doing Right Now:
We are buying two servers and not clustering them so we essentially lose the resources of the second server as it's not being used for anything else.
Why would you assume you need/want to cluster them? If you don't have licensing for that, that is yet another expense to make that work.
Also, what Hypervisor are they looking at? Hyper-V would be free in this case, but if they are a VMWare shop, then clustering is even more expense.
This is a 100% Hyper-V shop.
Isn't one of the strenghts of virtualization that you can allocate resources in a way where wasted resources are at the very least greatly reduced? An almost completely wasted server is the opposite of what we are trying to do in the modern world especially when their reasoning is hardware redundancy. You can have hardware redundancy with clustering and retain all of your resources. Seems like a no brainer to me.
I think you are thinking of cloud computing. Hyper-V might be able to do that, but again, you still need licensing that supports the moving of Windows Based VMs between hosts. And that's not how many SMBs are.
-
@dafyre said in What Are You Doing Right Now:
@wirestyle22 said in What Are You Doing Right Now:
@dafyre said in What Are You Doing Right Now:
@wirestyle22 said in What Are You Doing Right Now:
@Dashrender The server is on it's way out anyway so it solves two problems but I'm really put out.
Were you actually in that meeting, or did just you just happen to hear it as you walked by?
My desk is directly next to their office so I hear everything
Perhaps you should mention it to them... "By the way, I heard you talking about raid 5..." and then explain to them why it's a bad idea?
Government, it being a bad idea was probably the point.
-
@wirestyle22 said in What Are You Doing Right Now:
@Dashrender said in What Are You Doing Right Now:
@wirestyle22 said in What Are You Doing Right Now:
We are buying two servers and not clustering them so we essentially lose the resources of the second server as it's not being used for anything else.
Why would you assume you need/want to cluster them? If you don't have licensing for that, that is yet another expense to make that work.
Also, what Hypervisor are they looking at? Hyper-V would be free in this case, but if they are a VMWare shop, then clustering is even more expense.
This is a 100% Hyper-V shop.
Isn't one of the strenghts of virtualization that you can allocate resources in a way where wasted resources are at the very least greatly reduced? An almost completely wasted server is the opposite of what we are trying to do in the modern world especially when their reasoning is hardware redundancy. You can have hardware redundancy with clustering and retain all of your resources. Seems like a no brainer to me.
The goal of virtualization is abstraction.
-
@scottalanmiller said in What Are You Doing Right Now:
@wirestyle22 said in What Are You Doing Right Now:
@Dashrender said in What Are You Doing Right Now:
@wirestyle22 said in What Are You Doing Right Now:
We are buying two servers and not clustering them so we essentially lose the resources of the second server as it's not being used for anything else.
Why would you assume you need/want to cluster them? If you don't have licensing for that, that is yet another expense to make that work.
Also, what Hypervisor are they looking at? Hyper-V would be free in this case, but if they are a VMWare shop, then clustering is even more expense.
This is a 100% Hyper-V shop.
Isn't one of the strenghts of virtualization that you can allocate resources in a way where wasted resources are at the very least greatly reduced? An almost completely wasted server is the opposite of what we are trying to do in the modern world especially when their reasoning is hardware redundancy. You can have hardware redundancy with clustering and retain all of your resources. Seems like a no brainer to me.
The goal of virtualization is abstraction.
With that comes the ability to allocate resources in a way that is much more cost effective as well though correct?
-
@wirestyle22 said in What Are You Doing Right Now:
@dashrender If they aren't considering all of this I highly doubt they have calculated iops too. I don't get triggered a lot but I feel like we're setting ourselves up to fail.
Your company definitely is. You should mention these things.
They need to take a step back and do something like - run DPACK for a few weeks, use that information to build the servers that will be in the summer, and they will probably be overkill for the DAG situation.
-
@Dashrender said in What Are You Doing Right Now:
@wirestyle22 said in What Are You Doing Right Now:
@dashrender If they aren't considering all of this I highly doubt they have calculated iops too. I don't get triggered a lot but I feel like we're setting ourselves up to fail.
Your company definitely is. You should mention these things.
They need to take a step back and do something like - run DPACK for a few weeks, use that information to build the servers that will be in the summer, and they will probably be overkill for the DAG situation.
I don't know how these guys are going to react to me involving myself in this process but I'd rather deal with social consequences than IT consequences
-
@wirestyle22 said in What Are You Doing Right Now:
@scottalanmiller said in What Are You Doing Right Now:
@wirestyle22 said in What Are You Doing Right Now:
@Dashrender said in What Are You Doing Right Now:
@wirestyle22 said in What Are You Doing Right Now:
We are buying two servers and not clustering them so we essentially lose the resources of the second server as it's not being used for anything else.
Why would you assume you need/want to cluster them? If you don't have licensing for that, that is yet another expense to make that work.
Also, what Hypervisor are they looking at? Hyper-V would be free in this case, but if they are a VMWare shop, then clustering is even more expense.
This is a 100% Hyper-V shop.
Isn't one of the strenghts of virtualization that you can allocate resources in a way where wasted resources are at the very least greatly reduced? An almost completely wasted server is the opposite of what we are trying to do in the modern world especially when their reasoning is hardware redundancy. You can have hardware redundancy with clustering and retain all of your resources. Seems like a no brainer to me.
The goal of virtualization is abstraction.
With that comes the ability to allocate resources in a way that is much more cost effective as well though correct?
In some cases, not in others.
