Example Thread Gone Hilariously Wrong
-
@scottalanmiller said in Step Back When Asking Questions:
@DustinB3403 your post did eventually get approved.
Oh good. I thought I f-bombed the guy and it was being held up by that jackass over there. . .
-
Here is the thread, as I suspect the OP will request it deleted as there is no way to look like anything but an idiot when all is said and done...
Post 0 - OP
Which Hyper-V DR/resilience option should I use?I am running 2 Windows Server 2016 hosts in different locations connected via a WAN. VM1, running a bespoke application, is on Host A (the 'primary server') and is accessed externally by 'customers' via an external IP address. There is a requirement that in the event of failure of Host A and therefore VM1, Host B (the 'backup server') will take over automatically. What should I be looking to implement to fulfill this requirement?
Edited Jan 12, 2018 at 8:36 AM
Post 1
Scott Alan Miller Jan 12, 2018 at 8:13 AM
symonay wrote:
I am running 2 Windows Server 2016 hosts in different locations connected via a WAN.
Your title is about Hyper-V. But now you said Windows Server. Which one is it completely changes what the question is. We will need clarification.
Post 2
Scott Alan Miller Jan 12, 2018 at 8:19 AM
You have two different questions going on, one is how to have a DR environment for either Windows Server or Hyper-V. If you are using Windows Server, of course you'd never have DR since running Windows Server as a physical install is not a production installation that would ever qualify for DR (or installation in the first place.) The answer, in that case, would obviously be "get Hyper-V", then ask again.
If you really have Hyper-V and Windows Server was a typo, then we have to ask... what is the application? You can't ask these things in a bubble, it doesn't work that way. The answer for Active Directory or SQL Server are very different than the answer for a static web application. So even having Hyper-V isn't enough, on its own, to tell you how your DR will be designed. There are generic answers that people will act like are fine to give you, but they are not reliable depending on the workload. So without more details, we can't actually answer. People probably will either in the hopes of a quick vote for best answer, or just because they don't understand DR for various workloads. But any answer that doesn't require more information is just wrong.
The second part is you are asking how to use two different IP addresses. That's a completely different question. And doesn't really make sense to be combined here. Why would you have different IP addresses at the same time you were dealing with a server failure? This doesn't make sense. The servers and the IPs fail indepedently. So tying them together like ths would create failures and complexity and fix nothing. It is guaranteed there is a fundamental misunderstandnig by whoever told you to look into this and they don't understand what they are looking for.
You should step back and talk about your goals. You are too much "in the weeds" in this situation to be sure what you are looking for.
Post 2
symonay Jan 12, 2018 at 8:23 AM
Scott Alan Miller wrote:Your title is about Hyper-V. But now you said Windows Server. Which one is it completely changes what the question is. We will need clarification.
I have re-worded to try and provide clarification.
Post 3
Scott Alan Miller Jan 12, 2018 at 8:28 AM
symonay wrote:
I have re-worded to try and provide clarification.
Same issue, I don't see what has changed. Hyper-V in the title; Windows Server in the question.
Post 4
symonay Jan 12, 2018 at 8:30 AM
Scott Alan Miller wrote:Same issue, I don't see what has changed. Hyper-V in the title; Windows Server in the question.
Please look again. Hyper-V is running on Windows Server 2016 hosts.
Post 5
Scott Alan Miller Jan 12, 2018 at 8:31 AM
The only people that can answer the DR question for sure are the writers of the bespoke application. They and they alone know how the application fully works and can tell you how its state is handled, what state it has, and how different DR scenarios will impact it. They may know that it is stateless and that any platform DR (Hyper-V based DR) will work just fine for it. But it would be one odd bespoke application if that were the case. More likely they have data states that need to be handled and, in most cases, this is handled by the bespoke developers writing a DR mechanism into the application or utilizing a DR feature in the database that they chose to use.
But when dealing with bespoke apps, we can't look it up or know it based on how the app works, since we know nothing about it.
Post 6
Scott Alan Miller Jan 12, 2018 at 8:33 AM
symonay wrote:
Please look again. Hyper-V is running on Windows Server 2016 hosts.
Ah, well two problems there...
Never once does it say that. It says the question is about Hyper-V, then states that you have Windows Server hosts. It's one or the other.
Hyper-V cannot run on Windows Server hosts, only the other way around. So there is both an issue with what you wrote, and an issue with what you think that you have.
Note: At this point the OP has been edited multiple times. We don't know how many edits are made, only that it is still being edited. This is the break point where we first know that the words "VM" are added to the OP. But the OP does not tell us that they are added, and we only know of them after we've finally worked out that he is, in fact, virtualizing.
Post 7
BBigford Jan 12, 2018 at 8:33 AM
symonay wrote:I am running 2 Windows Server 2016 hosts in different locations connected via a WAN. VM1, running a bespoke application, is on Server A (the 'primary server') and is accessed externally by 'customers' via an external IP address A. There is a requirement that in the event of failure of Server A and therefore VM1, Server B (the 'backup server') will take over automatically and traffic will flow internally over the WAN to Server B. What should I be looking to implement to fulfill this requirement?
You would want a site to site VPN with very good throughput if you're going to use Quick or Live Migration to fail over automatically. Quick Migration means they aren't in a cluster, Live means they are. They are very similar, and often get used interchangeably, but not the same. This would be a "hot" state, meaning it's ready to go.
If you can't get good throughput, use VM Replication. This is setup per VM, and you just right click on it to set it up. VM Replication requires manually turning the fail over VM on, but works great with low throughput. This is a "cold" state. It replicates every so often, but isn't just ready to go during a failure, you have to turn it on.
Each has its pros and cons (Live Migration being automatic but requires more throughput which can be spendy; VM Replication which can cost less in throughput, but requires manual intervention which might be 10 minutes or less of down time).
Both are features in Hyper-V. Just depends on your budget for throughout and acceptable level of downtime.
Post 8
Scott Alan Miller Jan 12, 2018 at 8:34 AM
After clarification, this is what you have. It very clearly does not state what you think that it states. Title and description are conflicting and don't reflect what you thought you were asking.
Screenshot of Conversation Was Here
Your hosts can only be one or the other. If they are Windows, they aren't Hyper-V.
Post 9
Scott Alan Miller Jan 12, 2018 at 8:36 AM
Here is some information on Hyper-V basics that will help you to understand what you have today. Hyper-V is a type one hypervisor, that means it runs only on bare metal. What you have, we presume, is the anti-pattern of using the Hyper-V as Windows Server role installation method. This doesn't change what Hyper-V is, where it runs, or that Windows is a VM and not running on the host; but it is not recommended as it cripples Hyper-V in many ways.
Post 10
Scott Alan Miller Jan 12, 2018 at 8:38 AM
BBigford wrote:
You would want a site to site VPN with very good throughput if you're going to use Quick or Live Migration to fail over automatically. Quick Migration means they aren't in a cluster, Live means they are. They are very similar, and often get used interchangeably, but not the same.
If you can't get good throughput, use VM Replication. This is setup per VM, and you just right click on it to set it up. VM Replication requires manually turning the fail over VM on, but works great with low throughput.
Each has its pros and cons (Live Migration being automatic but requires more throughput which can be spendy; VM Replication which can cost less in throughput, but requires manual intervention which might be 10 minutes or less of down time).
Both are features in Hyper-V.
That only works if the workload can be handled that way. It might, but it might not.
Notation: At this point the OP did his last, as recorded here, edit of his original post but was not telling anyone so that people looking later were seeing much more details and clarification than was there at the time of the discussion. And as the edits were secret, no one knew to check back.
Post 11
Scott Alan Miller Jan 12, 2018 at 8:40 AM
What is needed regardless of how the failover piece happens, is something that detects the failure and manages the failover. You can do this using something like a load balance (F5, HA-Proxy, etc.) - but without knowing the application we can't answer what can be used here. This can failover very quickly but is expensive and you need the load balancers in a place that won't fail, or the issue returns.
Or you can do this with DNS, and just change to which location the traffic is pointed. This only works if requests are done via URL rather than IP, and will require a detection mechanism, automation of the DNS updates, and time for DNS changes to propogate before clients will see the system come back online.
Post 12
Scott Alan Miller Jan 12, 2018 at 8:41 AM
As a bespoke application, the application itself might have mechanisms for this completely built into it. For example, it might use a client that can do this detection and failover. That would be rare, but with a bespoke app, anything is possible.
Post 13
symonay Jan 12, 2018 at 8:44 AM
Thank you BBigford. Scott Alan Miller, I'm sure you have the best of intentions, but you don't come across well; if your intentions are other than altruistic then I don't know why you bother. You seem hell bent on picking holes and trying to make me feel stupid. There are 2 install types for Hyper-V, headless on Windows Core or with GUI i.e. Window Server, so I don't believe my terminology is that far out i.e. I have Hyper-V running on a Windows Server 2016 installation. You just seem to be being pedantic for the sake of it.
And there we have it, that's where the OP left it. After pointing out the mistakes over and over. After stating what he stated wasn't possible. After explaining it. Provided an education video on the exact topic, he then rants about it and gets ALL details of Hyper-V wrong - both how you can install it as well as once again stating that he had Hyper-V on top of Windows even though we'd covered, over and over again, that that wasn't possible.
Clearly, regardless of what he knew before starting this thread, it turned into a thread he wasn't able to follow. In the end, we don't know if he has Hyper-V, we don't know if he has Windows, we don't know how the application works... he has nothing useful from this except a chance to look like an idiot.
-
@scottalanmiller here is a Picture of the conversation
-
Sadly, this turned into way too much about one thread, lol. In the end, I can't believe anyone is actually this stupid. This HAS to be pure trolling. I mean seriously. Four times pointing out that he was getting something dead simple wrong... but he stuck to his guns looking dumber and dumber with each post.
-
I ended up reporting the OP for trolling and the guy supporting him for unprofessional "demanding condescension" behaviour that doesn't fit with the community.
I'm pretty sure that the thread was a set up. People get their rocks off on this stuff.
-
And the topic is locked
-
@dustinb3403 said in Example Thread Gone Hilariously Wrong:
And the topic is locked
Yeah, THREE HOURS after I reported two trolls in it.
-
To be completely honest I've been moderated and muted faster than this topic...
-
What's great is I warned, in the very first post, about how people were going to behave and how fake answers were going to be given if he didn't give the full information. And look, exactly what happened.
-
@dustinb3403 said in Example Thread Gone Hilariously Wrong:
To be completely honest I've been moderated and muted faster than this topic...
I know, right? I reported this one and wanted it shut down a long time ago. Honestly, the OP is, I think, a newbie and should be censored. It was totally inappropriate to run a "set up" thread.
-
You are still having your posts allowed through in a trickle.
-
And the op just asked to have the topic deleted
-
@scottalanmiller said in Example Thread Gone Hilariously Wrong:
You are still having your posts allowed through in a trickle.
Yeah because I'm being moderated as if I'm @RojoLoco lol
-
@dustinb3403 said in Example Thread Gone Hilariously Wrong:
And the op just asked to have the topic deleted
Exactly as predicted.
-
Being stupid usually is unpleasant..
-
Yeah, imagine how unpleasant it was for the people who were trying to be helpful but he chose to set up and attack for being so.
His whole "making me feel stupid" thing is just BS. Another person can't really make you feel stupid, only you can make yourself feel stupid. What a line to use.
-
Had he just been willing to admit he was confused and either ask for help, or read what was provided to him, it could have been a useful, educational thread. Instead, he decided to take offense to not knowing everything and lash out at others for his own shortcomings. He could have handled that well, quite easily, and looked very normal as it is a completely normal thing to be confused about.
-
-
@sam-is-pedantic said in Example Thread Gone Hilariously Wrong:
@scottalanmiller
https://www.psychologytoday.com/blog/in-love-and-war/201311/why-do-we-blame-victimsGreat article and a good example of what happened.
-
This paragraph sums it up well: "Victim blaming is not just about avoiding culpability—it's also about avoiding vulnerability. The more innocent a victim, the more threatening they are. Victims threaten our sense that the world is a safe and moral place, where good things happen to good people and bad things happen to bad people. When bad things happen to good people, it implies that no one is safe, that no matter how good we are, we too could be vulnerable. The idea that misfortune can be random, striking anyone at any time, is a terrifying thought, and yet we are faced every day with evidence that it may be true."
That, and I assume that lots of people feel that they have technical gaps and prefer to feel coddled than to feel that they should be exposed. Humans react negatively to the pressure to improve ourselves, which comes from identifying flaws, gaps, mistakes, etc.
