KVM Backing and Support
-
@scottalanmiller said in I think I am missing something about Hyper-V....?:
KVM has solid backing, and solid support options.
@scottalanmiller Can you fork this into a new thread?
-
Indeed we can.
-
:folded_hands: :thumbs_up:
-
@scottalanmiller Can you elaborate on your statement "KVM has solid backing, and solid support options."
-
@fateknollogee said in KVM Backing and Support:
@scottalanmiller Can you elaborate on your statement "KVM has solid backing, and solid support options."
KVM has solid, dedicated vendors behind it. Like Red Hat, Suse, Canonical, IBM, Amazon, Digital Ocean, etc. Companies that have invested heavily and aren't going anywhere. KVM took over from Xen to be the cloud leader, as well. So it has the most mature and robust user base, as well (their customers are the biggest of the big customers.)
KVM is primarily provided through support companies - companies whose bread and butter is providing support for systems; instead of software companies that primarily sell licenses and are not known for their ability to execute reliable support. Vendors like RH, Suse, Canonical, and loads of HV vendors and similar, provide primary KVM support.
In the wild, in the third party support market, you have "fewer" KVM support resources, but the ration of good to bad makes it just as easy or realistically easier to get viable KVM support. Hyper-V and VMware suffer from the problem of everyone and their nephew claiming to know and support them so the quality support people, who are definitely out there, are silenced in a sea of ineptitude. Finding good support, no matter how abundant it is, can be a struggle because you have to identify and filter out hundreds of unskilled GUI button clickers who don't know the systems at all before you come across a true support person and then you have to not filter them out on accident.
-
Would you consider Scale & Nutanix to also be KVM "vendors"?
-
It would be nice to have a list of "quality" 3rd party KVM support providers.
-
@fateknollogee said in KVM Backing and Support:
Would you consider Scale & Nutanix to also be KVM "vendors"?
They are. They package and support KVM.
-
@fateknollogee said in KVM Backing and Support:
It would be nice to have a list of "quality" 3rd party KVM support providers.
That would be like listing the support providers for Linux. There are thousands and thousands of them.
-
My biggest questions with KVM still revolve around backup solutions for it.
-
@dustinb3403 said in KVM Backing and Support:
My biggest questions with KVM still revolve around backup solutions for it.
What's the question? Use agent based, or modern DevOps backups like you should use in most environments regardless. Some KVM environments have agentless backups, but how often is that a good thing rather than a crutch or just a marketing thing?
-
But it's so convenient and easy to be able to back up (agentless) VMs at the hypervisor level with the ability to restore files within VMs like you can with Hyper-V backup solutions.
-
@obsolesce said in KVM Backing and Support:
But it's so convenient and easy to be able to back up (agentless) VMs at the hypervisor level with the ability to restore files within VMs like you can with Hyper-V backup solutions.
VMWare and Hyper-V solutions both have this.
-
@scottalanmiller said in KVM Backing and Support:
Use agent based,
Screw that shit. Let's just jump back to 1999 shall we?
@scottalanmiller said in KVM Backing and Support:
modern DevOps backups
This is not even a thing. DevOps is not backups. If you are trying to talk about stateless, that does not apply to the SMB almost 100%, and still doens't account for backing up the data bits.
-
@jaredbusch said in KVM Backing and Support:
This is not even a thing. DevOps is not backups. If you are trying to talk about stateless, that does not apply to the SMB almost 100%, and still doens't account for backing up the data bits.
"We use <Insert cloud buzzword> so we don't need Backups" has been the rallying cry of developers who think they understand infrastructure for the past 10 years. Fun thing I learned over lunch with a CIO last week, business continuity insurance will reject your claim if this was the reason (Too many people have thought using AWS alone was a backup).
There are other reasons for doing Hypervisor based backups API's for backup and replication beyond being able to restore a few files.
-
CBT. Having an API that does differential block WITHOUT having to read the data. Not all OS's Applications have a clean way to do this. Doing block based agent backups with 20TB of data means I have to read 20TB of data to find the 500MB that changed. Long backup windows, and if you use a log structured back end for your primary storage a brutal random IO storm.
-
Test/Dev workflows that use the backup for copying out clones for test/dev (See what Cohesity, Rubrick, and Veeam Labs enable).
-
Write spilter API's. VAIO lets you get to a block based few second RPO without using a kernel module that can bring down the host if it fails.
-
Data analytics off secondary copy. Commvault can do this, and I expect to see more data mining done against secondary storage copies.
-
Function as a part of a larger orchestration suite for testing DR (SRM as an example).
-
-
@obsolesce said in KVM Backing and Support:
But it's so convenient and easy to be able to back up (agentless) VMs at the hypervisor level with the ability to restore files within VMs like you can with Hyper-V backup solutions.
- Is it? How is "so convenient" really important in IT? Unless you can put a dollar value on that convenience, it's not relevant.
- It comes at a cost, a cost of reliability and performance. I see loads of shops getting useless backups because they thought convenience trumped "working". It encourages lazy, bad backups and processes.
- Once you do all the due diligence and effort to get good backups the difference in effort between agentless and agent is generally nominal.
-
@jaredbusch said in KVM Backing and Support:
@scottalanmiller said in KVM Backing and Support:
Use agent based,
Screw that shit. Let's just jump back to 1999 shall we?
@scottalanmiller said in KVM Backing and Support:
modern DevOps backups
This is not even a thing. DevOps is not backups. If you are trying to talk about stateless, that does not apply to the SMB almost 100%, and still doens't account for backing up the data bits.
DevOps is not backups, but good DevOps changes how backups need to be used. Backups are no longer necessary to make up for system administration deficiencies. Backups in the non-DevOps world are often used as a crutch, an expensive one.
-
@jaredbusch said in KVM Backing and Support:
@scottalanmiller said in KVM Backing and Support:
Use agent based,
Screw that shit. Let's just jump back to 1999 shall we?
It's not a jump back, it's sticking with the more enterprise solution. Agentless is limited in scope and requires support at the hypervisor, OS, and application level. Essentially no enterprise shop can use it, as there is no agentless system that supports the range of apps that shops use. So no enterprise has moved to agentless. Many use it as an "extra" piece, making backups more complex and more expensive, rather than less.
Really, for the time being, agentless is mostly just marketing hype. So jumping to "tried and true" rather than "sounds impressive and is rarely thought through" is exactly what we should want.
-
@storageninja said in KVM Backing and Support:
"We use <Insert cloud buzzword> so we don't need Backups" has been the rallying cry of developers who think they understand infrastructure for the past 10 years.
Actually, no one says that. No one.
The point is smaller, faster, more focused backups of relevant data. Not loads of fluff to sell more hardware.
-
@scottalanmiller said in KVM Backing and Support:
Actually, no one says that. No one.
The point is smaller, faster, more focused backups of relevant data. Not loads of fluff to sell more hardware.I've had this conversation with a shocking amount of developers who thought Cloud=automated unlimited backup and DR.
It often flows from "Everything's in git hub!" (follow by littering of state all over the place, or ignoring that just because the database was delivered by PaaS, or deployed from a template doesn't mean you don't need to do SOMETHING to protect the data.
