BRRABill's Field Report With XenServer
-
@BRRABill said in BRRABill's Field Report With XenServer:
@scottalanmiller said
But how is it a lot more work? You've said many times XS is a breeze to install. Fix the array (which probably means recreating it from scratch if the drives have failed), reinstall XS, restore the VMs. I don't see how the USB saves you time here.
That XS is easy to install isn't the issue, that's beside the point. It's dealing with the array, which is not easy to recreate, that is the issue.
Installing XS after a failure is silly, have it ready to go before a failure.
Restoring an array when you've lost the array controller is a big deal, I can't understand why you'd even consider opting to have this in your process. This could easily be what kills you and causes data loss. It's time consuming, complex and a lot of risk. For what purpose?
Basically, you are looking at going against the advice of every hypervisor vendor and the industry which are recommend for a reason and doing something only nominal advantageous (saving what, $10?) but... why? You are trying to downplay the advantages, but you are failing to explain why "just a little worse" isn't still "worse."
-
Do a Pros and Cons list. Pros to using SD / USB are solid. Maybe not epic, but they are there. Cons are... what? What factors are driving you to want to question a nearly universal industry standard from both the IT and the vendor sides?
Not that questioning is not good, but industry accepted best practices normally exist for extremely strong reasons. Reinvesting the wheel or approaching things from a "I must be a special case" are basically always wrong.
-
@scottalanmiller said
Basically, you are looking at going against the advice of every hypervisor vendor and the industry which are recommend for a reason and doing something only nominal advantageous (saving what, $10?) but... why? You are trying to downplay the advantages, but you are failing to explain why "just a little worse" isn't still "worse."
Hey I am questioning everything!
I'm in agreement it is the right way to do it. I just don't understand why the other way is so bad. This is purely educational at this point.
Other than if the XS installation gets hosed. Unless some update goes bad, what could kill XS on the array?
-
@scottalanmiller said
Not that questioning is not good, but industry accepted best practices normally exist for extremely strong reasons. Reinvesting the wheel or approaching things from a "I must be a special case" are basically always wrong.
Purely educational.
-
@BRRABill said in BRRABill's Field Report With XenServer:
@scottalanmiller said
Basically, you are looking at going against the advice of every hypervisor vendor and the industry which are recommend for a reason and doing something only nominal advantageous (saving what, $10?) but... why? You are trying to downplay the advantages, but you are failing to explain why "just a little worse" isn't still "worse."
Hey I am questioning everything!
I'm in agreement it is the right way to do it. I just don't understand why the other way is so bad. This is purely educational at this point.
Other than if the XS installation gets hosed. Unless some update goes bad, what could kill XS on the array?
The array controller dies... too many disks die...
-
@dafyre said
The array controller dies... too many disks die...
Right but if that happens (or you lose more than 1 disk which has happened to me) your data is all gone anyway, right? ANd you are restoring the VMs from backup anyway.
-
@BRRABill said in BRRABill's Field Report With XenServer:
I'm in agreement it is the right way to do it. I just don't understand why the other way is so bad. This is purely educational at this point.
Mistake: You changed "not as good" to "so bad". That's confusing you.
-
@scottalanmiller said
Mistake: You changed "not as good" to "so bad". That's confusing you.
You know, that is a great way of looking at it.
-
@BRRABill said in BRRABill's Field Report With XenServer:
@dafyre said
The array controller dies... too many disks die...
Right but if that happens (or you lose more than 1 disk which has happened to me) your data is all gone anyway, right? ANd you are restoring the VMs from backup anyway.
So in SOME cases, it's not AS advantageous, but still advantageous, right? You are saying that while it is better, it's not enough better to justify not doing something worse? That makes no sense.
-
@BRRABill said in BRRABill's Field Report With XenServer:
@scottalanmiller said
Mistake: You changed "not as good" to "so bad". That's confusing you.
You know, that is a great way of looking at it.
I've gotten used to that problem with RAID 5. Someone will argue with me that RAID 5 isn't "that bad" and they'll forget that they are only trying to minimize the worse, not show why it's okay.
Because only the "best" option should ever be considered.
In the case of local boot, unless we have a reason that it is "better", then it is "worse", so avoid it.
-
And to be clear, there is no reinstall of XS on an existing drive, right?
It trashes whatever you install it on?
-
@BRRABill said in BRRABill's Field Report With XenServer:
And to be clear, there is no reinstall of XS on an existing drive, right?
It trashes whatever you install it on?
Right. Thankfully, it is really easy to make an exact copy of the XenServer drive. Hrm, I actually need to do that tody. I'll try to take screenshots and do a how-to type writup.
-
-
@DustinB3403 said in BRRABill's Field Report With XenServer:
See here
http://mangolassi.it/topic/8537/how-to-clone-a-xen-usb-on-windows
Well, that's one less thing for me to document here Then again, I'm going to be doing this on a live, running XenServer (LVM is great, you should sing it's praises.)
-
For me, it's just another way of doing things the way I am used to doing them, instead of using the easy way already built in.
I can take my server down for a bit if I need to, so just shut down the VM, copy it to my test XS setup, redo my array, install XS to USB, and copy the VM back. Easy, and all built-in.
Why try to do something any harder?
-
@BRRABill said in BRRABill's Field Report With XenServer:
For me, it's just another way of doing things the way I am used to doing them, instead of using the easy way already built in.
I can take my server down for a bit if I need to, so just shut down the VM, copy it to my test XS setup, redo my array, install XS to USB, and copy the VM back. Easy, and all built-in.
Why try to do something any harder?
Because it isn't harder, it's easier. none of the steps that you mention are easy in a failure condition.
-
@scottalanmiller said
Because it isn't harder, it's easier. none of the steps that you mention are easy in a failure condition.
I meant moving my install from the array to USB. But you are right in your comment as well.
-
When you copy/export a VM, does it also copy removable storage you have attached?
-
@BRRABill said
When you copy/export a VM, does it also copy removable storage you have attached?
Well, I figured this one out for myself. Yes, yes it does.
I have my Tandberg drive hooked up to this one VM I am trying to copy. I kept wondering why it needed so much space to copy. Then I removed the Tandberg from the list of VDs, and voila, the space needed was way down.
So that's kind of wierd. Does it convert the removable into permanent in the copy?
-
@BRRABill I doubt that it converts it from removable to "internal" but it would have to create the backup with everything as attached.
Otherwise what good is it?