Burned by Eschewing Best Practices
-
@stacksofplates said in Burned by Eschewing Best Practices:
@scottalanmiller said in Burned by Eschewing Best Practices:
Nothing there implies that it is bad. Rarely used and bad are different kinds of concepts.
Again, I never said that. But they aren't really. I rarely use my truck, because it's bad. They can be different, but not always.
They absolutely are different. That you can find a situation where one leads to the other does not imply that they are connected. Any example where they are not connected proves that. That's how connection works. Like being old makes you a woman. Sure many old people are women, but it's not being a woman that makes you old.
-
FreeNAS is a little like RAID 5 for Winchester disks. There are times when it is a disaster, there are other times that it is the "second best choice." The sad reality for RAID 5 on Winchester drives today is that it is never the best choice - this means that no matter how good it is, it's never the first choice and therefore we would never consider it at all. Even as the second choice, once we know it is never above a second choice, we simply rule it out from the beginning and take a critical view of anyone who chooses it because it means that they left it in the decision pool for no reason.
FreeNAS would be a good choice if there was no FreeBSD. Of course, if there wasn't FreeBSD, FreeNAS would not exist, but imagine that it did. A BSD OS with ZFS and amazing stability, it would be pretty awesome and basically totally replace FreeBSD in the storage pool options if only FreeBSD wasn't there.
But it is the nature of the bad idea of trying to make it as an appliance that takes it to a place always behind FreeBSD or TrueOS so that it never "bubbles to the top." No need to entertain it, it never is choice number one. But if number one didn't exist, it would be basically every time that FreeBSD or TrueOS is today.
-
So in the same vein as RAID 5 (on Winchesters on new arrays) the issue is not that FreeNAS can't be "good enough" on its own, but we know ahead of time with essentially no effort that it will never be the best choice. So no matter how close it gets to second place in some crazy situations, the issue is not the final choice but rather the decision making process that would lead someone to consider it in the first place. It's a failure of quality decision making - why even entertain something that is known to not be the best choice? Good decision making would rule it out before you begin so that confusion does not arise as humans tend to be thrown off by a surplus of options. And the idea of "good enough" often confuses people.
-
Wow this thread blew up from last night.
Anyways I agree with @stacksofplates, you have repeatedly come across as saying that FreeNAS is a bad product (regardless if you meant idea).
Because of the lack of support, the community preaching and the other reasons.
So what makes OF worse (or a never use appliance) in comparison to FreeNAS, since that was the original question. What makes OF worse than FreeNAS?
-
@DustinB3403 said in Burned by Eschewing Best Practices:
Wow this thread blew up from last night.
Anyways I agree with @stacksofplates, you have repeatedly come across as saying that FreeNAS is a bad product (regardless if you meant idea).
Because of the lack of support, the community preaching and the other reasons.
So what makes OF worse (or a never use appliance) in comparison to FreeNAS, since that was the original question. What makes OF worse than FreeNAS?
What makes OpenFiler worse is that it's broken - or at least that's what I recall hearing. It has a flaw that can just outright lose your data.
-
@DustinB3403 said in Burned by Eschewing Best Practices:
Anyways I agree with @stacksofplates, you have repeatedly come across as saying that FreeNAS is a bad product (regardless if you meant idea).
Because of the lack of support, the community preaching and the other reasons.
Can you find where I have said that? This has been repeated by other people over and over again (putting words in my mouth) but I've never been shown where I supposedly said it. I've been super clear that the idea is bad and that FreeNAS is among the best examples of this bad idea (the best of the options, I mean.) The community is an example of why the idea is bad, because storage experts would never use something in this ilk. But the community is not the product nor the vendor.
Please provide a reference. You personally say that I say this all the time, whether you are joking or mean to, and I correct you constantly. If I've actually said that FreeNAS itself is bad instead of FreeNAS' community being bad, it's users being confused or the idea being bad, I need that reference.
-
@scottalanmiller Read what I said scott.
I said you come across as saying FreeNAS is a bad product, and in the never use category when there are other options such as CentOS or RHEL.
The reference was already provided by @stacksofplates in the screenshots. The wording is very easy to interpret as "OMG FreeNAS is satan!"
-
@DustinB3403 said in Burned by Eschewing Best Practices:
So what makes OF worse (or a never use appliance) in comparison to FreeNAS, since that was the original question. What makes OF worse than FreeNAS?
It's not even related.
https://community.spiceworks.com/topic/373443-why-we-recommend-against-openfiler
The issues with OpenFiler are unrelated to the idea of them both being bad.
- OpenFiler is a dead product, no work on it in ~6 years. FreeNAS is alive and well.
- OpenFiler was only ever a one man basement project. FreeNAS has a large hardware vendor behind it.
- OpenFiler is built on a dead OS and hasn't been patched in eons. FreeNAS is based on FreeBSD which is very much alive.
- OpenFiler was fundamentally unstable and would die at random. FreeNAS is stable.
- OpenFiler threatens people who expose them. FreeNAS has nothing to expose.
- OpenFiler has a web interface with no function as it doesn't work. FreeNAS has a working web interface.
- OpenFiler is officially listed as unusable by VMware. FreeNAS is fine with VMware.
- OpenFiler has a known broken iSCSI stack. FreeNAS has a working iSCSI stack.
- OpenFiler made their money selling fixes to their intentionally broken system. FreeNAS has nothing to sell except support.
- OpenFiler support is questionable if it even exists (we tried.) FreeNAS has a large, known support team.
- Both products suffer from being in the Jurassic Park Effect of being a product category that doesn't make sense as you can use their underlying OS better than you can use the packaged product.
That's a lot of differences.
-
@Dashrender said in Burned by Eschewing Best Practices:
What makes OpenFiler worse is that it's broken - or at least that's what I recall hearing. It has a flaw that can just outright lose your data.
Multiple flaws like that, and that's only part of the issue. And what is even worse is that the flaws are known, there is a fix and you only get the fix if you buy it. Paying to have an intentionally included flaw fixed is similar to extortion. It's not just paying for patches, it's paying for patches when the bugs are intentional.
-
@DustinB3403 said in Burned by Eschewing Best Practices:
The reference was already provided by @stacksofplates in the screenshots. The wording is very easy to interpret as "OMG FreeNAS is satan!"
So you admit that I never said it and it is people just making up implications that are not there. The screenshots very clearly say nothing of the sort. The feeling that I have said this is caused by you and others repeating this all the time, then when there is a something that says that idea is bad, people just assume that I meant the product. It's you that has been saying it, not me.
-
@DustinB3403 said in Burned by Eschewing Best Practices:
I said you come across as saying FreeNAS is a bad product, and in the never use category when there are other options such as CentOS or RHEL.
I said it is in a never use category. Find where I said it was a bad product. And don't try to conflate the two as being related as we clearly know that they are not.
The examples that have been provided are only of the later, which I have never disputed.
-
The lumping together of OpenFiler and FreeNAS that people contribute to you Scott I'm sure at this point was a misunderstanding.
But you rattle off a dozen reasons why NAS appliances (software, not hardware) are just a bad idea - and people, those of us who are not you, walk away with the assumption that they are all just bad.
Right or wrong, it's how people feel when they walk away from one of your conversations about it.
-
@Dashrender said in Burned by Eschewing Best Practices:
But you rattle off a dozen reasons why NAS appliances (software, not hardware) are just a bad idea - and people, those of us who are not you, walk away with the assumption that they are all just bad.
A bad idea. But I've been told that I explicitly said that FreeNAS itself is bad above and beyond being a bad idea or having a bad community or having people use it improperly. That I've said it's a bad idea is not in question. That it is a bad product other than that seems to be things that other people have said but keep getting attributed to me. Which happens with a lot of things, all the time. This is a regular problem of "Scott said..." but no one is sure where I said it, but tons of people repeat it. Same thing happened with the moderation issue on SW a few weeks ago, it turned out eventually that it was something that someone else said that I said, in a place that was private that I knew nothing about, and everyone kept repeating that so that I never found out exactly what people even thought that I had said. It's becoming a very common thing. RAID 5 info being repeated with only partial context is another that is treated like that.
-
@Dashrender said in Burned by Eschewing Best Practices:
Right or wrong, it's how people feel when they walk away from one of your conversations about it.
Well of course, most people work on implication instead of what is actually said. That's why marketing works so well, people assume that something is going to be said and even if the marketing doesn't say it, they just assume it to be true. There is no way to fix that.
-
@scottalanmiller said in Burned by Eschewing Best Practices:
@DustinB3403 said in Burned by Eschewing Best Practices:
I said you come across as saying FreeNAS is a bad product, and in the never use category when there are other options such as CentOS or RHEL.
I said it is in a never use category. Find where I said it was a bad product.
It is very easy to correlate never use = bad product
Even if never use =! bad product
Because if you would never use it, it must be bad in comparison to the alternatives.
-
Here are some thought experiments:
RAID 5 should never be used on a new array of Winchester drives today, given current market available drives and those available since ~2009. Does this imply that RAID 5 itself is bad?
RAID 5 is the only single parity RAID approach that should be used today, surpassing RAID 4 which likewise replaced RAID 3. Does this mean that RAID 3 and 4 are bad, or were bad? Or does it mean that something has surpassed them? What does bad mean in this context?
CentOS 4 is old and deprecated. You should never deploy it. Does this mean that CentOS 4 is bad?
The point here is that in some ways, all of these things can be determined to be good. RAID 5 remains the best RAID for broad use on SSD arrays today and was the top pick for Winchester arrays twenty years ago. Does losing popularity or efficacy in one arena make something that was good, and did not change, bad?
RAID 3 and 4 were the best options, for many things, in their time. Does the invention of something more robust change them from good to bad, even though they themselves did not change?
CentOS 4 was new long ago. Does the fact that it has been replaced and has aged make it bad? Is bad and old the same thing?
Likewise, or similar, FreeNAS itself is a good product, but one that is in a category that makes no sense and should never be used. Does the fact that FreeBSD is better make FreeNAS bad? Is bad an absolute issue or a relative one? Does something turn from good to bad through no action of its own?
-
@DustinB3403 said in Burned by Eschewing Best Practices:
It is very easy to correlate never use = bad product
But it should not happen. That's an implication that does not exist in the statement. You should never use the "almost as good, second best product". But that doesn't make it bad, just not the best.
- Best is always determined by a variety of factors and is relative to the situation.
-
@DustinB3403 said in Burned by Eschewing Best Practices:
Because if you would never use it, it must be bad in comparison to the alternatives.
It must be worse compared to, not bad compared to. And it is, FreeNAS is worse compared to FreeBSD, but better compared to nearly everything else. But it only takes being worse than one to be "Never use."
-
Here is a great example. There is an Olympic runner (he's immortal, bear with me) and he runs in every summer Olympics for a century. Over that time, he never, ever wins the gold. But he wins the silver every, single time. He was never the best, ever. Not once, but we would not call him bad for being the runner up consistently, right? If you wanted the fastest runner, he's never the best bet, he's always fast but never on top.
FreeNAS and RAID 5 (under the Winchester new array conditions) are like that. Sometimes they are second best for a scenario but never first. Because of good decision making, we rule out a known secondary player because we don't want to accidentally consider something we know isn't the right answer to make picking the right answer easier and to not waste time on something we know can't be right. But that doesn't mean that it is at the bottom of the heap or bad, only that it is not number one.
-
@scottalanmiller said in Burned by Eschewing Best Practices:
@DustinB3403 said in Burned by Eschewing Best Practices:
Because if you would never use it, it must be bad in comparison to the alternatives.
It must be worse compared to, not bad compared to. And it is, FreeNAS is worse compared to FreeBSD, but better compared to nearly everything else. But it only takes being worse than one to be "Never use."
Do you see how one could read this statement as FreeNAS is worse than FreeBSD but better than RHEL.
?
