You are right, it was mostly just a bunch of testing work (done by us) to validate that modern Linux kernels can handle the increased ARC usage in a bunch of different scenarios. That limitation was put in place a LONG time ago, and since then the reasons for it aren't necessarily valid anymore. We place a premium on stability and didn't want to just flip that switch without knowing it was going to be reliable long term. The only caveat is that if you run KVM / VMs, you want to make sure you leave enough room for ARC to do ARC things, without squeezing it too much. We'd recommend adjusting it down in that use-case, otherwise you are fine.@@jttech44
@@jttech44 you can set it to 100gb for example but its not flexible based on usage, especially with multiple VMs going on/off, which changes system debt on ram, much better to be dynamic and change with needs of the server, and use every bit of ram you paid for. the current way is a workaround at best.
It’s a shame what’s happening to FreeBSD but I will hold onto my CORE installation until I’m 6 feet under or burned into ash. I choose TrueNAS because of its BSD beckend. I want my stability and reduced headaches over troubleshooting (if any).
Your comment about ceph is one I have thought about a lot wrt truenas. I have petascale ceph deployments and it's been rock solid for many years. high availability through version upgrades, hardware upgrades, firmware upgrades, no corruption, incredibly customizable and scalable.. honestly it's changed my life in that way
Thanks Tom, for bringing clarity here. I have a few systems running on glusterFS and was surprised when I found out Kubernetes 1.25 dropped support, also due to abandoned project components as far as I understood. Feels like everyone is moving either to the enterprisey Ceph or Suse/Rancher's Longhorn, for a simpler solution. I've always been very happy with Gluster. Sad to see them disappear.
It's not really surprising that glusterFS is slowly going away due to alternatives like CEPH. There is also CEPHfs. CEPH got a good track record for support and keeping things stable. Now I have to wonder about CEPH's future due to IBM buyout. For now they're keeping CEPH free but that could change.
@@Darkk6969 CEPH is great... until it's not. That's the problem. It gets very very complex to admin very quickly if you go outside of the "everyone does it like this" bubble. That's not a problem in and of itself, but, many don't know that going in and wind up either ditching it or becoming the 'CEPH guy'. I really do wish they'd make something akin to a CEPH-lite, that just gave you the basics with plenty of guardrails to keep the newbies from doing something that only technically works but will probably lead to disaster.
@@jttech44exactly my thoughts. I am using ceph, too but as you said: if something is getting out of hand it's getting very complex very quickly. Longhorn might fill that gap. Seems a bit rough around the edges sometimes but overall I'm starting to like it in the newest versions.
I think Apple not contributing back to the project is such a shame. Part of me thinks they want it to die so the can effectively close source the project and bring it in to Apple's walled garden.
@@James-ln6li Open-source driver (not binary) should be easier to move to something like BSD, even if it has to be rewritten. Licensing might be the bigger issue (GPL--->BSD).
@@brodriguez11000 No.. drivers are by nature tied to the operating system 's interfaces on which they are running. But Nowadays drivers are not that important.. what is important is unrestricted availability of full documentation of the device. Vendors have moved their "custom magic" (or IP whatever they call it these days) to firmware or firmware like loadable modules..
i am loving SMB and especially ISCSI in core for data that is needed right away. Very stable and just works, faster than scale too. VMs and apps go on my unraid servers and unraid even though not as fast , I can set all my drives to sleep every 45min if not used
I have been spinning down my hard disks for years and have had no problems with them. And YES, if you live in Europe the costs for energy are significantly higher, and it can be worthwhile to reduce consumption. If all the HDDs I have running didn't go into spindown, I would pay about €70 more a year. If the whole thing goes on like this for a few more years, I can easily afford 1-2 new HDDs as replacements and I'll still be more economical. But I'm only talking about my case here, and in my setup it's worth it.
@@DanFromBlackForest your experience is on a small set of drives. Statistical data on significant numbers of drives - as well as physics - prove that spinup/spindown cycles are not good for hdds.
@@BoraHorzaGobuchul absolutely right, he just lucky, numbers dont lie, especially in larger environments, its commonly known that spindown is not good but someone who never experienced it wont learn. at the end of the day we cant teaching people that are arrogant enough to ignore the facts because they never experienced it and wont believe what others say
Yes, iXsystems wrote that at least two features would be dropped with 13.1. So be it. I doubt if I will rebuild everything to work on SCALE after having spent so much work on CORE.
Thanks for this Tom. I do occasionally try Scale on a spare machine, knowing we'll be pushed that way if that's where the development is occurring. My last experience was an app install that simply would not un-install. I suppose though, I'm part of the problem because I didn't submit a ticket!
I first setup my NAS in mid 2011, running FreeNAS. It's the same hardware, excluding the disks, which original were 8 x 3TB WD RED's, but now are 8 x 10TB WD GOLD's. Besides that everything else is the original stuff. Never had a major issue with upgrading the OS, been an absolutely diamond, love it. It instilled in me a great love of FreeBSD. I'm sad that SCALE has taken the limelight, and I am concerned what will become of CORE, and FreeBSD. Such a fine OS, and IMO way more stable than Linux based OS's. It's a doddle to upgrade between releases. Do hope we don't see it fade into obscurity.
Part of the problem with BSD is the Sheer arrogance of the BSD engineers and experts that provide very little support to the new players or even appear to care about the rest of the world of IT . They act like they are UNIX Gods, Too many times me going to a BSD sight to get help you get these high and mighty experts telling you " read the hand book" And when you do, you will find that it lacks real clarity or explanation to assist you with your specific issue. All this arrogance only causes the would be interested parties to go to something that may be not better but easier to resolve issues because of the abundance of support and documentation. BSD engineers need lessons in humility to help increase the base.
I tried moving to scale from core but it wouldn't install any of the plug-ins it kept saying processing or something like that it's been awhile the rest of the system worked great looked great etc just couldn't get any programs to install at all. So I went back to core and thats where I've been at
Was using core for a long time when I first started out with truenas but as you said the Apps and Jails were atrocious and constanly having issues for me. Switched to scale and have never looked back. Runs all my apps and storage so much better than core ever did.
I'll be on Core until the sun expands and swallows the Earth. And maybe for a little while after that. Nothing against Scale... I've just been on Core since it was FreeNAS, it meets my needs, and I see no reason to rip it out and replace everything.
Same here. Same installation upgraded all the way from FreeNAS, rock solid, does its job perfectly, no bug, very stable I can't ask more of it. No point moving elsewhere tbf.
few months ago I installed my TrueNAS server at home. And I was thinking long which one to choose from. Eventually went with Scale as I know Linux very well unlike FreeBSD.
It isn't even that FreeBSD is bad or difficult or anything of the sort, it's just so so similar to linux that it's frustrating to use be cause it's juuuuust different enough to throw you off.
Thanks for the timely video Tom! And thanks to Kris Moore for the forum posts. I'm a home user spinning up my first build and was leaning Core since i just need a simple file server and no apps or vms or containers. But there's some pressure to just go with Scale on Day 1 and be done with it.
I thought that (and hardware compatibility) but chose Core for its outright stability. I do have a secondary with Scale on it but prefer the stability of Core on BSD. Scale UI can be nicer (not always but mostly).
I will eventually move from Core to Scale on my long time running TrueNAS server. Just needs to be tested and mature a bit more. FreeBSD has been rock solid for me but Linux got much bigger support.
Do open source jobs development jobs actually get advertised outside of their communities? I don't recall ever seeing one on a jobsite. Might be UK thing
Not supporting Docker is an absolute dealbreaker for me, so Scale all the way if I'm going with TrueNAS. I will never not have Docker on a server again.
Scale does not give the same reliability as Core. That's a point. Users choose Core explicitly because of its stability. Forcing users to an unstable alternative will bring people also to alternatives outside the ixsystem world. So they have to be VERY careful
Correct, remember we aren't talking days or months here, but still YEARS of support for CORE on our side. Lots of time for things to mature :) @@Darkk6969
If you were doing a new deployment, would you lean more toward Core or Scale? I ask because I'm currently in the process of ordering TrueNAS machines from iXsystems at work. I'd planned to go with Core simply because we don't need any of the additional features in Scale but I've had some lurking concerns about the trajectory of BSD. My biggest concern with Scale was ARC but it sounds like improvements there are right around the corner. Unrelated - iXsystems is wonderful to work with. We liked them so much that we're ordering all the servers for our current project from them.
For Enterprise both are options, but Enterprise customers tend to gravitate towards the more conservative things. You can keep running CORE/Enterprise for a while and if you want to jump to SCALE Enterprise at some later point we can handle that as well, already doing that for customers as they ask for it. You are not locked into one or the other.
@@LAWRENCESYSTEMS I'd be interested to know why exactly you hold that opinion. I hear it said alot, but, never qualified, and, I can't really think of anyone more qualified than you. I went the opposite direction last time I deployed a ton of storage, because it seemed that hardware support was much better on scale out of the box, which simplified things.
As much as I love FreeBSD it's fairly easy to see Scale is just more viable for the future. I'm still running Core but I've finally gotten around to learning scale (again, first time was a bit of a disaster) since it doesn't take a genius to see the writing on the wall. That said I'll run Core for as long as it's supported.
wouldn't Core just stick around as long as ixsystem still uses FreeBSD for TrueNAS Enterprise since it's basically the same OS with a couple of features stripped out?
Scale is where you should put your money. Now, let's see if Netgate will do the same with pfSense (don't talk to me about tsnr (Linux-based) that is wasting precious developers time into something no one really cares or use - instead for example of adding features to pfSense like a central management console for lots of pfSense, a real IPS/IDS, EDR, etc., and turn it into a true security device (aka NGFW) on the network for modern needs, not just a periphery firewall for home labs).
Only an insane person would want to support two different platforms for their solution. They will drop core eventually, there's no new feature work on it. In the meantime, it remains the more stable platform of the two within it's more limited scope.
The shift in FreeBSD support is way bigger than just TrueNAS’es future. It affects any other current FreeBSD project, all other BSD systems as a whole, and to an extent, Linux projects as well. It shows people are shifting from creativity and flexibility, into commodity and longevity. Which aren’t bad things per se. But less folks dealing with more complex elements in favor of a vendor dictated solution is. Long term, that will become a problem: as more and more Linux drivers and components are getting updates from big companies than from the developers in the community, those with no ties to a company would also disappear. And as these same big companies keep adding traditional Linux companies into their portfolio, some only to later pull out the plug - RedHat being the last one to retire -, a lot of hardware will too fall short of support as there are no incentives for these same companies to support products not making them any money. Soon, long gone would be legacy hardware, and TCO would raise for everyone that could otherwise keep an old yet fully functional hardware - see Windows 11 official supported requirements and how many incompatible W10 machines are out there - operational for much longer.
Right. FreeBSD, the poster child for creativity and flexibility. You tried to use a USB 3.0 thumb drive to install? LOLs. Everyone knows USB 2 is the future, get with the times! And by vendor flexibility you mean the licensing no doubt. I ascribe to the law that states the bigger the company, the worse life is for everyone involved in the project known as living, but to hear a BSD fanboy decry the involvement of big companies when the very license itself was meant as a mating call to get more big companies to come and use it is the most funny thing I have read in 2024. Everyone has been totally free, like beer free man, to use BSD for anything they want. And yet, here we all are.
I'm still on the GlusterFS mailing list. Gluster had promise, but unfortunately, they took away one of the most useful features in version 6 (deprecated in version 5) where I couldn't create nor use a RAMdrive Gluster volume, and then export that to my IB network via NFSoRDMA. As a result, Gluster then had functionally, no use for me, so I stopped using it. If they had kept it in, it would've been worthwhile to keep it. *sigh...*
@@pantoqwerty are their no jails in scale or you just have to reset up using scale.specific versions. My main use case is Plex server. I don't mind setting that up again but if Plex wasn't available then that's an issue for me.
1:1 migration isn't possible, nor would you really want it to be. In terms of ZFS itself, yes, the pools import just fine. ZFS is ZFS, it isn't exotic. The config though, it's specific to what you're running, so you get to come up with a strategy for Scale if you're currently on Core. It isn't miles different, just, different.
you can do it from System -> Update then change the "Train" to a Scale version. I dont believe you can go back though. you **may** be able to just boot into a previous environment from the System -> Boot screen
We've been running Proxmox in a business environment for about a year now, I've been very happy and found it very stable. Development seems to be ongoing and healthy. This is just for a small company so we only have a couple of servers and not a huge need for complex features that vSphere/vCenter had that Proxmox might not yet, but it does everything we need.
Its sad to see BSD go this way, but its hard to see a different future. Maybe some of the devs can be persuaded to apostatise but bring some BSD rigour and methods across.
Well, the gluster removal kills my TrueNAS dev project. We were running a 6-node TrueNAS cluster in beta, waiting on the full production version. Bummer.
Clustering with Linux seems to be the same story as many other features and implementations - one particular project gets traction and popularity, may even get rolled into the mainstream kernel, but the project loses its user base or its vendor funded support and someone comes up with some completely different alternative and everyone abandons option A for option B. Without a strong vendor championing it, such as The Linux Foundation itself, we don’t see the longevity that we need. Changing technologies every few years is painful and wasteful.
Is Apple the 800 lb gorilla in BSD world? Does Apple depend on Darwin/BSD? Does Apple contribute to BSD? Same questions for Oracle (which purchased SUN Microsystems)? If Apple left BSD what would be left? Real time systems? Is there any software that needs BSD Kernel as opposed to Linux Kernel?
Let's hope enough people care.. Scale is overcomplicated and ineffiecient, being based on the ungodly mess that is modern day linux/debian. Hoping for a fork and proper forward /ongoing develop TN-Core using BSD based technologies. Hearing that Core13.x will be supported for a while, but it's planned to be the last BSD version. IX systems have betrayed the BSD community bigtime here.
The iPhone and Android captured all the kids, and we just don't have enough new blood to keep both the BSDs and Linux up to speed with all the technology churn. Be realistic. There was no betrayal at all.
Ceph would be cool but it’s so hard to get a highly performant ceph system, for something like running VM’s off of it You need a ton of nodes and it also doesn’t support RDMA natively so it ends up missing out on a lot
I personally am glad I moved away from truenas, their eco system is to buggy, they introduce breaking changes, they don’t follow common standards ie file system permissions, container management/kubernetes switched to unraid and it took me 3 days to do what took me over a month in truenas and I haven’t had ti go back and redo my configurations when I update my containers and I can go on
That's what happens when projects use a cuck license. Sad to see this happening, but FreeBSD was always the best choice of companies out of all of the BSDs.
Well, “backwaters” are safe. “Moving fast” isn’t a core value for any important backend software storing your assets. Slow is good, an epistemological value in itself. 😅👍
@@musiqtee Slow is not a virtue in technology. The pace and amount of resources working on Linux is an asset which is why it is the defacto operating system for the internet. BSD just can't compete... and I like OpenBSD... but it's at a standstill by comparison and that's not good for the things that matter: performance, support, features, and security.
Thanks for the video Tom! One note: ZFS ARC Fix for SCALE is landing in Dragonfish BETA.1 - ETA next week 2024/2/6
Awesome! Looking forward to it!
What exactly is it fixing? Aside from not using nearly as much memory for the ARC as it could, which is configurable, it seems to work.... fine?
The ears of an entire community just perked up.
You are right, it was mostly just a bunch of testing work (done by us) to validate that modern Linux kernels can handle the increased ARC usage in a bunch of different scenarios. That limitation was put in place a LONG time ago, and since then the reasons for it aren't necessarily valid anymore. We place a premium on stability and didn't want to just flip that switch without knowing it was going to be reliable long term. The only caveat is that if you run KVM / VMs, you want to make sure you leave enough room for ARC to do ARC things, without squeezing it too much. We'd recommend adjusting it down in that use-case, otherwise you are fine.@@jttech44
@@jttech44 you can set it to 100gb for example but its not flexible based on usage, especially with multiple VMs going on/off, which changes system debt on ram, much better to be dynamic and change with needs of the server, and use every bit of ram you paid for. the current way is a workaround at best.
It’s a shame what’s happening to FreeBSD but I will hold onto my CORE installation until I’m 6 feet under or burned into ash. I choose TrueNAS because of its BSD beckend. I want my stability and reduced headaches over troubleshooting (if any).
Your comment about ceph is one I have thought about a lot wrt truenas. I have petascale ceph deployments and it's been rock solid for many years. high availability through version upgrades, hardware upgrades, firmware upgrades, no corruption, incredibly customizable and scalable.. honestly it's changed my life in that way
Thanks Tom, for bringing clarity here. I have a few systems running on glusterFS and was surprised when I found out Kubernetes 1.25 dropped support, also due to abandoned project components as far as I understood. Feels like everyone is moving either to the enterprisey Ceph or Suse/Rancher's Longhorn, for a simpler solution. I've always been very happy with Gluster. Sad to see them disappear.
It's not really surprising that glusterFS is slowly going away due to alternatives like CEPH. There is also CEPHfs. CEPH got a good track record for support and keeping things stable. Now I have to wonder about CEPH's future due to IBM buyout. For now they're keeping CEPH free but that could change.
@@Darkk6969 CEPH is great... until it's not. That's the problem. It gets very very complex to admin very quickly if you go outside of the "everyone does it like this" bubble. That's not a problem in and of itself, but, many don't know that going in and wind up either ditching it or becoming the 'CEPH guy'. I really do wish they'd make something akin to a CEPH-lite, that just gave you the basics with plenty of guardrails to keep the newbies from doing something that only technically works but will probably lead to disaster.
@@jttech44 Yep. That is one of the reasons why I am using ZFS on my ProxMox servers.
@@jttech44exactly my thoughts. I am using ceph, too but as you said: if something is getting out of hand it's getting very complex very quickly. Longhorn might fill that gap. Seems a bit rough around the edges sometimes but overall I'm starting to like it in the newest versions.
Its a shame to see BSD go this way. Companies to contribute back should be Apple and Sony.
I think Apple not contributing back to the project is such a shame. Part of me thinks they want it to die so the can effectively close source the project and bring it in to Apple's walled garden.
Drivers should be able to move between the operating systems.
@@brodriguez11000 what?
@@James-ln6li Open-source driver (not binary) should be easier to move to something like BSD, even if it has to be rewritten. Licensing might be the bigger issue (GPL--->BSD).
@@brodriguez11000 No.. drivers are by nature tied to the operating system 's interfaces on which they are running.
But Nowadays drivers are not that important.. what is important is unrestricted availability of full documentation of the device.
Vendors have moved their "custom magic" (or IP whatever they call it these days) to firmware or firmware like loadable modules..
i am loving SMB and especially ISCSI in core for data that is needed right away. Very stable and just works, faster than scale too. VMs and apps go on my unraid servers and unraid even though not as fast , I can set all my drives to sleep every 45min if not used
What's the point of spinning down drives? With all the empirical evidence showing that having them running 24/7 is actually better for them?
@@BoraHorzaGobuchul yes, spindown is stupid, someday he will learn it the hard way, just let it be
I have been spinning down my hard disks for years and have had no problems with them. And YES, if you live in Europe the costs for energy are significantly higher, and it can be worthwhile to reduce consumption. If all the HDDs I have running didn't go into spindown, I would pay about €70 more a year. If the whole thing goes on like this for a few more years, I can easily afford 1-2 new HDDs as replacements and I'll still be more economical.
But I'm only talking about my case here, and in my setup it's worth it.
@@DanFromBlackForest your experience is on a small set of drives. Statistical data on significant numbers of drives - as well as physics - prove that spinup/spindown cycles are not good for hdds.
@@BoraHorzaGobuchul absolutely right, he just lucky, numbers dont lie, especially in larger environments, its commonly known that spindown is not good but someone who never experienced it wont learn. at the end of the day we cant teaching people that are arrogant enough to ignore the facts because they never experienced it and wont believe what others say
Yes, iXsystems wrote that at least two features would be dropped with 13.1.
So be it. I doubt if I will rebuild everything to work on SCALE after having spent so much work on CORE.
Not worth it, go to unraid; trust me
Thanks for this Tom. I do occasionally try Scale on a spare machine, knowing we'll be pushed that way if that's where the development is occurring. My last experience was an app install that simply would not un-install. I suppose though, I'm part of the problem because I didn't submit a ticket!
Excellent coverage as always, Tom! Many thanks.
Your channel is very informative. As a Free/TrueNAS user for a very long time. I love your coverage.
I first setup my NAS in mid 2011, running FreeNAS. It's the same hardware, excluding the disks, which original were 8 x 3TB WD RED's, but now are 8 x 10TB WD GOLD's. Besides that everything else is the original stuff. Never had a major issue with upgrading the OS, been an absolutely diamond, love it. It instilled in me a great love of FreeBSD. I'm sad that SCALE has taken the limelight, and I am concerned what will become of CORE, and FreeBSD. Such a fine OS, and IMO way more stable than Linux based OS's. It's a doddle to upgrade between releases. Do hope we don't see it fade into obscurity.
Part of the problem with BSD is the Sheer arrogance of the BSD engineers and experts that provide very little support to the new players or even appear to care about the rest of the world of IT . They act like they are UNIX Gods, Too many times me going to a BSD sight to get help you get these high and mighty experts telling you " read the hand book" And when you do, you will find that it lacks real clarity or explanation to assist you with your specific issue. All this arrogance only causes the would be interested parties to go to something that may be not better but easier to resolve issues because of the abundance of support and documentation. BSD engineers need lessons in humility to help increase the base.
Jeez and i thought Linux neckbeards were bad
I tried moving to scale from core but it wouldn't install any of the plug-ins it kept saying processing or something like that it's been awhile the rest of the system worked great looked great etc just couldn't get any programs to install at all. So I went back to core and thats where I've been at
Was using core for a long time when I first started out with truenas but as you said the Apps and Jails were atrocious and constanly having issues for me. Switched to scale and have never looked back. Runs all my apps and storage so much better than core ever did.
I'll be on Core until the sun expands and swallows the Earth. And maybe for a little while after that.
Nothing against Scale... I've just been on Core since it was FreeNAS, it meets my needs, and I see no reason to rip it out and replace everything.
Same here. Same installation upgraded all the way from FreeNAS, rock solid, does its job perfectly, no bug, very stable I can't ask more of it. No point moving elsewhere tbf.
Ditto. I hope Core remains supported for a long time.
I can't believe that! Looks very sad.
@@Hatch3dLabs there are no plans to replace core with scale, both have their justification for existence
few months ago I installed my TrueNAS server at home. And I was thinking long which one to choose from. Eventually went with Scale as I know Linux very well unlike FreeBSD.
It isn't even that FreeBSD is bad or difficult or anything of the sort, it's just so so similar to linux that it's frustrating to use be cause it's juuuuust different enough to throw you off.
@@jttech44 AMEN to that ;-)
Thanks for the timely video Tom! And thanks to Kris Moore for the forum posts.
I'm a home user spinning up my first build and was leaning Core since i just need a simple file server and no apps or vms or containers. But there's some pressure to just go with Scale on Day 1 and be done with it.
I thought that (and hardware compatibility) but chose Core for its outright stability. I do have a secondary with Scale on it but prefer the stability of Core on BSD. Scale UI can be nicer (not always but mostly).
I will eventually move from Core to Scale on my long time running TrueNAS server. Just needs to be tested and mature a bit more. FreeBSD has been rock solid for me but Linux got much bigger support.
Do open source jobs development jobs actually get advertised outside of their communities? I don't recall ever seeing one on a jobsite. Might be UK thing
THANKS! I just commented this yesterday on your other video APPRECIATE you making this video!!!!!
Not supporting Docker is an absolute dealbreaker for me, so Scale all the way if I'm going with TrueNAS. I will never not have Docker on a server again.
Scale does not give the same reliability as Core. That's a point. Users choose Core explicitly because of its stability. Forcing users to an unstable alternative will bring people also to alternatives outside the ixsystem world. So they have to be VERY careful
Core will be around for awhile. They're still testing and developing Scale.
Correct, remember we aren't talking days or months here, but still YEARS of support for CORE on our side. Lots of time for things to mature :) @@Darkk6969
If you were doing a new deployment, would you lean more toward Core or Scale? I ask because I'm currently in the process of ordering TrueNAS machines from iXsystems at work. I'd planned to go with Core simply because we don't need any of the additional features in Scale but I've had some lurking concerns about the trajectory of BSD. My biggest concern with Scale was ARC but it sounds like improvements there are right around the corner.
Unrelated - iXsystems is wonderful to work with. We liked them so much that we're ordering all the servers for our current project from them.
If just for storage, I would still go with core
For Enterprise both are options, but Enterprise customers tend to gravitate towards the more conservative things. You can keep running CORE/Enterprise for a while and if you want to jump to SCALE Enterprise at some later point we can handle that as well, already doing that for customers as they ask for it. You are not locked into one or the other.
@@LAWRENCESYSTEMS I'd be interested to know why exactly you hold that opinion. I hear it said alot, but, never qualified, and, I can't really think of anyone more qualified than you.
I went the opposite direction last time I deployed a ton of storage, because it seemed that hardware support was much better on scale out of the box, which simplified things.
@@jttech44 It works very well, is very stable and I know I can migrate later if needed.
As much as I love FreeBSD it's fairly easy to see Scale is just more viable for the future. I'm still running Core but I've finally gotten around to learning scale (again, first time was a bit of a disaster) since it doesn't take a genius to see the writing on the wall. That said I'll run Core for as long as it's supported.
wouldn't Core just stick around as long as ixsystem still uses FreeBSD for TrueNAS Enterprise since it's basically the same OS with a couple of features stripped out?
Scale is where you should put your money.
Now, let's see if Netgate will do the same with pfSense (don't talk to me about tsnr (Linux-based) that is wasting precious developers time into something no one really cares or use - instead for example of adding features to pfSense like a central management console for lots of pfSense, a real IPS/IDS, EDR, etc., and turn it into a true security device (aka NGFW) on the network for modern needs, not just a periphery firewall for home labs).
Not content related , but where to huy that Open source 24/7 t-shirt?
Also interested
Love them both, never had any issues.
Only an insane person would want to support two different platforms for their solution. They will drop core eventually, there's no new feature work on it. In the meantime, it remains the more stable platform of the two within it's more limited scope.
Thank you. Just needed a confirmation that i made the right choice. I think i did.
The shift in FreeBSD support is way bigger than just TrueNAS’es future. It affects any other current FreeBSD project, all other BSD systems as a whole, and to an extent, Linux projects as well. It shows people are shifting from creativity and flexibility, into commodity and longevity. Which aren’t bad things per se. But less folks dealing with more complex elements in favor of a vendor dictated solution is. Long term, that will become a problem: as more and more Linux drivers and components are getting updates from big companies than from the developers in the community, those with no ties to a company would also disappear. And as these same big companies keep adding traditional Linux companies into their portfolio, some only to later pull out the plug - RedHat being the last one to retire -, a lot of hardware will too fall short of support as there are no incentives for these same companies to support products not making them any money. Soon, long gone would be legacy hardware, and TCO would raise for everyone that could otherwise keep an old yet fully functional hardware - see Windows 11 official supported requirements and how many incompatible W10 machines are out there - operational for much longer.
Right. FreeBSD, the poster child for creativity and flexibility. You tried to use a USB 3.0 thumb drive to install? LOLs. Everyone knows USB 2 is the future, get with the times! And by vendor flexibility you mean the licensing no doubt. I ascribe to the law that states the bigger the company, the worse life is for everyone involved in the project known as living, but to hear a BSD fanboy decry the involvement of big companies when the very license itself was meant as a mating call to get more big companies to come and use it is the most funny thing I have read in 2024. Everyone has been totally free, like beer free man, to use BSD for anything they want. And yet, here we all are.
I'm still on the GlusterFS mailing list.
Gluster had promise, but unfortunately, they took away one of the most useful features in version 6 (deprecated in version 5) where I couldn't create nor use a RAMdrive Gluster volume, and then export that to my IB network via NFSoRDMA.
As a result, Gluster then had functionally, no use for me, so I stopped using it.
If they had kept it in, it would've been worthwhile to keep it.
*sigh...*
@Tom, is there or will there be an easy way to migrate to TrueNAS scale from Core should it ever become necessary?
I’ve been wondering whether you can boot up a server with the Scale image and import the config from Core. Jails won’t work, but is that an option?
@@pantoqwerty are their no jails in scale or you just have to reset up using scale.specific versions. My main use case is Plex server. I don't mind setting that up again but if Plex wasn't available then that's an issue for me.
1:1 migration isn't possible, nor would you really want it to be. In terms of ZFS itself, yes, the pools import just fine. ZFS is ZFS, it isn't exotic. The config though, it's specific to what you're running, so you get to come up with a strategy for Scale if you're currently on Core. It isn't miles different, just, different.
you can do it from System -> Update then change the "Train" to a Scale version. I dont believe you can go back though. you **may** be able to just boot into a previous environment from the System -> Boot screen
@@pantoqwerty if jails don't work is their another way to get Plex server running?
So, with the issues with VMware, the only option left is proxmox? Is there any news about the ongoing stability of proxmox?
We use XCP-NG th-cam.com/video/et54DxAC2uM/w-d-xo.html
We've been running Proxmox in a business environment for about a year now, I've been very happy and found it very stable. Development seems to be ongoing and healthy. This is just for a small company so we only have a couple of servers and not a huge need for complex features that vSphere/vCenter had that Proxmox might not yet, but it does everything we need.
Its sad to see BSD go this way, but its hard to see a different future. Maybe some of the devs can be persuaded to apostatise but bring some BSD rigour and methods across.
Will this downward trend of support for freebsd affect pfsense as well?
Yes
Sad but better to Focus on ONE Platform rather than two
Either way.... Its all fun and exciting 😎
Well, the gluster removal kills my TrueNAS dev project. We were running a 6-node TrueNAS cluster in beta, waiting on the full production version. Bummer.
"I'm not dead yet!" -TrueNAS Core
Clustering with Linux seems to be the same story as many other features and implementations - one particular project gets traction and popularity, may even get rolled into the mainstream kernel, but the project loses its user base or its vendor funded support and someone comes up with some completely different alternative and everyone abandons option A for option B. Without a strong vendor championing it, such as The Linux Foundation itself, we don’t see the longevity that we need. Changing technologies every few years is painful and wasteful.
Please please please don't kill Core.....! ❤
It is rock stable, just works perfect for storage.
RIP the LTT NAS config
Is it just me, or does just about everything that's OpenSource wind up this way? Seems like all sorts of projects are getting abandoned.
As someone who has worked in open source for over 20 years, yes. But this happens in closed source as well.
Is Apple the 800 lb gorilla in BSD world?
Does Apple depend on Darwin/BSD?
Does Apple contribute to BSD?
Same questions for Oracle (which purchased SUN Microsystems)?
If Apple left BSD what would be left? Real time systems?
Is there any software that needs BSD Kernel as opposed to Linux Kernel?
Let's hope enough people care.. Scale is overcomplicated and ineffiecient, being based on the ungodly mess that is modern day linux/debian. Hoping for a fork and proper forward /ongoing develop TN-Core using BSD based technologies. Hearing that Core13.x will be supported for a while, but it's planned to be the last BSD version. IX systems have betrayed the BSD community bigtime here.
The iPhone and Android captured all the kids, and we just don't have enough new blood to keep both the BSDs and Linux up to speed with all the technology churn. Be realistic. There was no betrayal at all.
Honestly I hated scale, if core goes away I don't know what I'll run but it wont be scale.
I mean, ZFS will be with us for the forseeable future, and runs on, well, basically anything worth running. Not terribly hard to roll your own.
My first 'Nas' I did everything manually on freebsd. Truenas makes it much, much easier/faster but it's completely possible to do without truenas.
Ceph would be cool but it’s so hard to get a highly performant ceph system, for something like running VM’s off of it
You need a ton of nodes and it also doesn’t support RDMA natively so it ends up missing out on a lot
May i know how to reduce zfs cach memory usage
On Scale? There are sysctl variables called something like zfs.arc.minsize and .maxsize which as the name implies control ARC sizing behavior.
Easy to google it, but don't. Buy more memory, use as much of it for ZFS as is possible.
I personally am glad I moved away from truenas, their eco system is to buggy, they introduce breaking changes, they don’t follow common standards ie file system permissions, container management/kubernetes
switched to unraid and it took me 3 days to do what took me over a month in truenas and I haven’t had ti go back and redo my configurations when I update my containers and I can go on
(Inhales)
Futue
In two TrueNAS videos I've basically learned nothing... Just a lot of "look here go here for info"
What's a "futue" ?
a typo
another abandoned google project lol, imagine counting on them for anything.
That's what happens when projects use a cuck license. Sad to see this happening, but FreeBSD was always the best choice of companies out of all of the BSDs.
Just dump Core and BSD. Linux is moving fast and BSD is a backwater.
TH-cam needs the laugh react … oh well here goes 😂😂😂😂😂
@@ashuggtubeok boomer
@@NetBandit70 hahaha I’m barely old enough to be Gen X … cope harder you neophyte 😂
Well, “backwaters” are safe. “Moving fast” isn’t a core value for any important backend software storing your assets. Slow is good, an epistemological value in itself. 😅👍
@@musiqtee Slow is not a virtue in technology. The pace and amount of resources working on Linux is an asset which is why it is the defacto operating system for the internet. BSD just can't compete... and I like OpenBSD... but it's at a standstill by comparison and that's not good for the things that matter: performance, support, features, and security.
I avoid Linux like the plague that it is
Thats a new one