Starting our next Open Source Project - TrueNAS SCALE

[IOSonic]

Thanks for the story and glad it's working for your family. Would love to hear tor views on what would improve the "cost of entry" (apart from complete docs which is on the list) ?

If it's working, please share your setup and application info.. its helps the developers and the community.

Hey @morganL . Before I answer your question, I'll point out that the primary function of a NAS is to store data reliably over a network; TrueNAS does this better than any other free & low-cost NAS software options--period. I'd recommend it to a small business for storage. It has enterprise features. And the interface makes it a great gateway for the ZFS novice (me).

TrueNAS could be better in its ancillary functions--specifically in its hypervisor & plugin system. My primary use for this system was serving docker applications backed by TrueNAS storage (think media server, backup applications, etc). Finding a lightweight Linux OS that would function properly as a docker host was extremely trying. As I recall, I had to rebuild RancherOS as an EFI image, and Arch and FedoraCore wouldn't boot at all. I eventually settled on Alpine, because that's what I could get to work . Even once it functioned, I ran into several weird issues, including abysmal network performance that was resolved with this advice (adding vlanhwtag as a system tunable). I was also unable to launch any jail in 12.0; I have a trunked, 4-port lag that carries a couple bridged vlans fashioned per KevDog's recommendation here, and the problem was related to that I believe. It self-resolved with 12.1.

I am leery that update will break something one day, and these FreeBSD bits being as esoteric as they are, I will be scouring forums looking for answers as I have needed to do often. For that reason, my permanent solution may be to eventually get some dedicated compute in the form of a pi cluster, but a native Kube system to replace all of the above complexity & fragility would be amazing. If it's built on debian, perhaps virtualbox could even succeed bhyve, allowing people to take advantage of its broad compatibility and many OVA images.

All of that aside, TrueNAS does plenty well, and I'll continue to use it and encourage others to do so.

P.S.
What's this all this language about "hyperconverged" and "scale-out" in your product page? I think there are some IT salespeople lurking in this fold.

----
Edited for clarity: replaced "iohyve" with "bhyve"

 

[KevDog]

@IOSonic -- I'm really excited by the scale project as well particularly given the K8s and KVM hypervisor, but I'm just curious why on the old system you just didn't install a Linux VM within bhyve and use this VM as your docker host? That's what I'm currently doing. I haven't run into that many problems with networking configuration or performance and the VLAN tagging -- yea that took awhile but it became rectified over time with further FreeNAS releases which kind of changed up things. I don't really find my setup all that fragile or prone to breaking.

 

[IOSonic]

Hey @KevDog, this is exactly what I ended up doing, but I ran into some problems along the way. I ended up using Alpine Linux for this purpose; it wasn't my first choice, but it was what I was able to get working. My desire for a separate computer cluster is informed by the trouble I ran into. Now that everything is sorted, it may run happily without problems; I'll wait and see. At any rate, thanks for all of your contributions to the forums. They helped a lot.

I'd missed that they plan to integrate KVM into Scale. Now that is exciting, even better than VB!

 

[ChrisRJ]

I am not sure what to make of many of the points discussed here. Personally, I would probably benefit from several of them. But so far I fail to see their value for enterprise customers. The latter have dedicated infrastructure for containers and VMs and would certainly not mix those functional areas with a NAS. Am I missing something here?

Thanks!

 

[HoneyBadger]

The latter have dedicated infrastructure for containers and VMs and would certainly not mix those functional areas with a NAS. Am I missing something here?

With apologies to @IOSonic for buzzwording, lots of enterprise customers are using "scale-out hyperconverged infrastructure" and combining compute+storage into a single building block.

 

[morganL]

I am not sure what to make of many of the points discussed here. Personally, I would probably benefit from several of them. But so far I fail to see their value for enterprise customers. The latter have dedicated infrastructure for containers and VMs and would certainly not mix those functional areas with a NAS. Am I missing something here?

There many types of Enterprise IT deployment; Global, DataCenter, Office, R&D, Cloud, Edge, Project, ShadowIT. TrueNAS provides a set of tools and users decide how to use them.

 

[morganL]

Hey @morganL . Before I answer your question, I'll point out that the primary function of a NAS is to store data reliably over a network; TrueNAS does this better than any other free & low-cost NAS software options--period. I'd recommend it to a small business for storage. It has enterprise features. And the interface makes it a great gateway for the ZFS novice (me).

TrueNAS could be better in its ancillary functions--specifically in its hypervisor & plugin system. My primary use for this system was serving docker applications backed by TrueNAS storage (think media server, backup applications, etc). Finding a lightweight Linux OS that would function properly as a docker host was extremely trying.

P.S.
What's this all this language about "hyperconverged" and "scale-out" in your product page? I think there are some IT salespeople lurking in this fold.

Hi, I think you helped answer some of the question with your own response. Yes, SCALE does have KVM, Docker/Kubernetes, and scale-out capabilities included. I'd call that "hyperconverged" eventually, but not quite now.

No question that its more complex initially, but as the software matures it will simplify deployments for many users. The need to set-up multiple machines with multiple OSes will reduce.. It should save people time and money. But nothing happens overnight.... its a journey. and glad to have your support.