Sounds like YouTube content. And unfortunately we have no control over that. It's like that TV show called "Jackass". Stupid stuff for stupid lulz. You won't find much of that around here because we tend to run that stuff out of the forums on a rail.
That seems on point.
I've never seen that in these forums except in the context of very high performance systems, or discussion of the enterprise systems that iXsystems sells as TrueNAS Enterprise. Most of the forum userbase are hobbyists or SOHO users, with the occasional IT guy charged with trying to build an alternative to TrueNAS Enterprise at a bargain price. A few of us do this professionally, but I don't see a lot of discussion of those systems.
Well. I didn't mean as literal as that. My point was, that someone who knows little to nothing about hardware, might find a gap between "I wanna use my laptop as a server" ($0) and "might we suggest you invest a little?" (>$0). I'm not saying that it's my impression, that people around here are unfair in their suggestions or are gate keepers - I'm very sorry, if it came of as that.
The name "FreeNAS" never meant that there were no costs involved. It meant that iXsystems would allow you to use their commercially developed enterprise quality software platform for free.
I think that's perhaps where some of the discretion (I don't even know, if that's the right word) comes from.
When I built my first server, I used a desktop motherboard, i7-2600S, with 8 GB of RAM and four laptop hard drives - just to test things out. Get a feel of FreeNAS and to see if that was an avenue I could follow.
At the time we had a QNAP 4-bay NAS and while it did serve us well, the 4 TB capacity and slow SMB performance was beginning to bog us down.
I wanted to see if I could built something myself and came across FreeNAS. I think it was right before version 11, because one of the very first things I did with the finished system was the update from 10 to 11.
This is the temp-system, still rocking the laptop hard drives but this time with 6 onboard SATA-connectors.
Finished the built in February/March '16 I think and a year later we moved from the apartment into our house.
This is the old TrueNAS box - spec. are in my sig.
Well, there are people here happy to answer questions.
This matches my experience as well. It's nice to see, people online on a "specialist forum" who actually helps out. As some of you might know, that isn't always the case on other forums.
That's not the first time this happened, it's a bit "too new" (and too expensive) for a solid setup, which is why "
the beard strokers" tend to wait for the bugs to be worked out and the drivers and such be really solid (
@jgreco is religious about this).
Well. It does make sense to have a kinda "if it ain't broke" and a somewhat conservativepersuasion when it comes to stuff like this. While it is fun to tinker, it isn't fun, when suddenly your data is gone - because, "raid is not a backup!" and I chose to pick up what-ever hardware other people (who know a lot more about this than me) couldn't get behind/support because they legitimately can't get behind it (for incompatibility reasons or other).
Right??? It's about 2011 vintage which explains the price, but what a solid setup! Tiny too. I'm not telling you what to do by any stretch, the thought is if you have a solid backup system (on the cheap) it's really quick/easy to isolate problems and test equipment while your main system is still running.
Well. I might go in that direction, but as someone else points out; that HTML5 iKVM is awesome. Right now, if the TrueNAS is borked or something, I can't even begin to explain all the hoops I have to go through because of the SuperMicro IPMI Java thingie.
It makes for popular and entertaining videos--which is his core business,
Trust me, I also mainly take it as being just that. I imagine I would feel the same, if he took vaseline to the front lens element of a cinema lens on a RED camera and ranted about it not being "pro" enough in a dark room, compared to a point and shoot under direct sunlight.
I just don't know enough about servers or production environments to know, when he is doing the equivalent to that.
And no (for anyone else reading); I'm not implying Linus (or any of his writers would even suggest that) but I also wouldn't hold it above him to do just that for the lulz... and to be honest. I would watch - for the entertainment. Not gonna lie.
(Once it's running, then migrate to TrueNAS, because an insane system deserves an insane OS.)
TrueNAS (and FreeNAS before that) was/is an incredible OS. Once you get through potential hardware issues, the "appliance"-thing does, well, apply. It has very much been "set and forget" for a large portion of the time for the current server - but that was also the mindset going in.
Not to be a "binary"-person, but sometimes I see it as either:
- Cheap/easy/accessible setup but continuous maintenance (updating, checking in, what-evs)
- Hard to setup / complicated but after that; minding its own business
Chose one.
or there's the guy who has an industrial-strength need and buys a high-end system and we drool over it, or out of nowhere a home user puts together a system with one-freekin'-hundred SSDs which is spectacularly nuts....uh, hold that thought....
This one is possibly going over my head. Are you implying something?
Those systems take the spotlight, but if you read the hardware guide Joe stresses saving as much money as possible and spending wisely, but frugally. That's the same thing I'm suggesting you do--other than the 100 SSDs--that's awesome!
To be honest, I was never going to save as much as possible - but I don't like throwing money out the window. If it is bonkers, but in any shape or form usable or practical - I'm in!
But. I imagine, even if I did strafe a little and - for instance - got the wrong HBA for this project, I can still use that HBA for other projects. It's not broken or anything.
Nope, nope, no, you're one of us. The transition is imperceptible to the person going through it, but it's an awesome transformation that imbues you with super-human ability and mad electronics skills. (That's how others will see you, anyway.) It's too late, you'll never get the addiction out of your head, you're one of us.
Now Get Back To Work, we're just as hooked on this as you are!
Well. I live in Denmark and as I have implied earlier, I'm a teacher. Our students had their finals this week and I'm not in any hurry. This is very much a side project (among a somewhat embarrassing long list of work in progress ideas). Other than that?
I'm on it, Sir!
It's weird, because every now and then you inadvertently impact someone in a way that pays off in spades.
I don’t know if you're pointing to an in-joke or something along those lines, just wanted to say, that I know what you mean.
A friend of mine is doing wildlife photography and he is very good at it. But as with many other things in life; sometimes you just want to shake it up a little. He wrote me on Messenger and aired his frustrations with needing new ideas. I had just recently gotten into astrophotography and showed him some images I had captured.
He was in and two weeks later we're both standing on a hill in the middle of nowhere, zipping beers and shooting stars.
You never know, what branch or twig, someone else hangs on to and it might take you by surprise.
As a teacher I sometimes experience a student - usually one of the quiet kids, who draws - that picks up on something and the next day, they show you something amazing they did. Drawings, DIY bluetooth speakers or that movie they saw, that they'll think you're also into. I love it!
X10 stuff is only marginally more expensive, but benefits from newer BMC firmware with the HTML5 iKVM. Plus the benefits of the newer generation, which are there, but not huge.
Both my servers are both "X10"s, but have different IPMI-solutions? How do I know, which to get if I follow WI_Hedgehog's advice and look into "last gen" stuff? Are that HTML5 iKVM thing even possible on older boards?
Learning is a humbling experience, no matter who you are.
Well. Some of us takes pride in liking to learn new stuff. Others, not so much - but they're very unlikely to be on these forums anyway, eh?
What an absolute treat to read along with the morning coffee!
I can definitely see traces of patience and habit of taking close notes of what works and what doesn't. I bet this has a lot to do with it:
Perhaps. Sometime, somewhere, I read: "The only difference between screwing around and science is writing it down." - something along those lines. Imagine my delight, when I found out who said it on camera. Anywho.
I've always been into doing that. I'm not very organized - shocker, I know - but I like to be systematic about things like this. Makes work much more effecient and helps to reveal patterns or point one in the right / other directions, when you've hit a road block.
It is quite possible it harks back the days when I did overclocking - but I think I remember I was into doing it before that. But, you're probably right, that doing overclocking helped cementing this very approach.
You've made a fantastic effort already!
Thanks. But a lot of it is also because of your guys helping me out - no doubt about it
At this point, I would grab the LSI SAS 9211-8i from your other rig, put one channel on each backplane.
I'd very much like to not do that. The 9211's are sitting in a 'live' system and I'd rather not dare to anger the "IT"-gods, by yanking hardware from a working system. You know: If it ain't broken. But I get what you're trying to have me do and I have considered it.
Right now, I'm more inclined to invest in another HBA - possibly that 9300-16i ting - but I'll have to do some more research.
At that stage, I'd remake a series of benchmarks ... which may not matter the slightest bit with 4 drives, but make a different story with 100 drives, and a budget PSU...
That's my plan as well. Originally I had planned it as such:
Filling / populating one backplane with SSDs, create a stripe, Rz1 or Rz2 and create a pool.
Fill another backplane, create a stripe/Rz1/Rz2, add vdev to pool.
And so on.
In that way, I'd be able to keep adding vdevs to the pool.
I'm not that concerned with redundancy for this project. It's nowhere near mission critical and I honestly had problems to come up with ideas for what I could do with the CrazyNAS - not unlike Linus' line in that video I linked to earlier: coming up with crazy ideas.
4c/8t do limit it in certain scenarios. But 128 GB RAM and 14TB combined (RAW) SSD storage capacity must have some applications.
... scratch disk for Adobe Premiere Pro?
I have no idea. If storage capacity was a concern, I'd sprung for some 4 TB SATA SSDs a long time ago. Currently they're priced at around $240-$300.
Just two of those are already more capacity than the entire CrazyNAS currently (right over 6 TB, with 58 drives).
So. If not for the storage capacity, then for the performance?
Well. Right now, it seems the old TrueNAS box is perhaps just as fast as the CrazyNAS - and I have no idea why. If it was RAM-speed I was testing (was using fio) the newer server should still be faster, right?
Funny, I did a project sharing the essence of this build a few years back. Gathering a much more modest compared to yours, 8-9 SSDs, and put them into an all SSD server, for basically the same reasons as you :)
Well. Let's see where I end up. We might both learn something then. You're more than welcome to inject / suggest ideas for use cases or benchmarks - and no, you don't have to justify them, other than what you'd have done with +100 SATA SSDs in a server
Also - this quote goes to my little hall of fame of fantastic quotes (in signature) :))
I'm starting to think, that I should've started collecting quotes from the beginning with this project
