Because the clone is a VMDK that resides on a VMWare box. Just start the VM.
There's your "restore." Back up and running in less than five minutes or so,
current to the last packet received. As I said, it's more like RAID 1 for
the server. If one of the drives in a mirrored set go out, do you restore
the bad drive, or replace it and return the machine to production?
It's not a restore. Certainly not in the traditional sense. Trying to call
it that only confuses things. Which is why were are standing on this dark
street corner having this conversation. Seriously: it's not called a
restore; it's called RTO or RTP (return to operation/production). It's not
just semantics; it's using the right word to describe what's going on. A cat
is not a dog. But call it what you like. Today is not a good day for me to
try and be polite. Makes no difference to me if you want to insist on
calling it the wrong thing.
-----Original Message-----
From: Ben Scott
Sent: Friday, September 13, 2013 5:04 PM
To: [email protected]
Subject: Re: [NTSysADM] In defense of image-based VM backups
On Fri, Sep 13, 2013 at 4:24 PM, Daniel Chenault <[email protected]>
wrote:
That's the point of the "hot clone" (I'm stealing that). It's an exact
image of the running server including any changes such as you describe
current
up to the last transmitted packet. There is no restore.
If there's no restore then why are you bothering cloning?
I'm not completely sure how to answer that. I don’t' think you understand
how the process works that I've already outlined.
You say there's no restore. I presume you are "hot cloning" to
*something*. Why bother doing that if you're never going to restore
it to anything?
Earlier you mention that it's cloning to a standby which you can
have up and running in four minutes or so. That's a restore, is it
not? The fact that your restore happens in four minutes doesn't mean
it's not a restore. It's just a really quick one.
And if you *are* going to restore it, then things get more
complicated. If the system was kind enough to crash cleanly, e.g.,
you unplugged the power cord accidentally, or the CPU just killed
itself and the system immediately bluescreen, then sure, restoring the
system as it was is great. It's a beautiful thing.
But what if the system got messed up by a software update, or
scrambled a filesystem, or had a malware compromise, or what-have-you.
Then you don't *want* to go back to the very last moment in time.
You want to go back to the last-known-good.
Then all the complications described previously still matter.
-- Ben
