Thanks for your insights, Tom! I personally haven’t encountered a motherboard failure yet, but this video will inform my future decisions. Syncthing and XCP-NG’s live migrations work so well that a motherboard failure can feel like a non-issue, as data would be transferred before the motherboard’s lifespan ends. Of course, there are unexpected failures, but in the use cases I’ve come across, latency or even MTTR wasn’t so critical that public cloud replication couldn’t bridge the gap. The more you know!
I would like to have a form of HA on my current hardware. It is very expensive for me to change the one I already have for ix hardware. I love TrueNas Scale but it would be nice to have a switch similar to the one from Proxmox. Thanks for the very informative videos. Greetings
I am still trying to understand how it's considered "highly available" solution. Let's imagine that the physical node itself burned out, literally or for example by a voltage spike. In a distributed solution, the data is physically distributed geographically, and the solution is at the software layer, no need to invest in high-end hardware. Have I missed something?
Would be nice huh, too bad it needs a license which is only available with their own systems. There are other solutions that use ZFS with dual port SAS etc that also have a free community version for non-production / home usage. You just need some more searching :-)
Apps should never be installed on a NAS, only the basic network share functions, for apps you have NUC where you install proxmox with some LXC and dockers. An App Store on a NAS is the biggest mistake they made
Thanks for your insights, Tom! I personally haven’t encountered a motherboard failure yet, but this video will inform my future decisions. Syncthing and XCP-NG’s live migrations work so well that a motherboard failure can feel like a non-issue, as data would be transferred before the motherboard’s lifespan ends. Of course, there are unexpected failures, but in the use cases I’ve come across, latency or even MTTR wasn’t so critical that public cloud replication couldn’t bridge the gap. The more you know!
I would LOVE a homelab level deployment of this. 😁
Fantastic content!
Thanks!
I would like to have a form of HA on my current hardware. It is very expensive for me to change the one I already have for ix hardware. I love TrueNas Scale but it would be nice to have a switch similar to the one from Proxmox. Thanks for the very informative videos.
Greetings
I am still trying to understand how it's considered "highly available" solution. Let's imagine that the physical node itself burned out, literally or for example by a voltage spike. In a distributed solution, the data is physically distributed geographically, and the solution is at the software layer, no need to invest in high-end hardware. Have I missed something?
Yes, this is not a cross datacenter distributed solution
So if sourcing the hardware correctly or just buying the similar Supermicro dual node server, could we get HA working?
Would be nice huh, too bad it needs a license which is only available with their own systems.
There are other solutions that use ZFS with dual port SAS etc that also have a free community version for non-production / home usage.
You just need some more searching :-)
What type of switches did you use for the data center deployment?
Probably some that support some form of MC-LAG, which allowes for LACP accross switches.
i think they use RSF-1 software for high availability
Apps should never be installed on a NAS, only the basic network share functions, for apps you have NUC where you install proxmox with some LXC and dockers. An App Store on a NAS is the biggest mistake they made
What is High availability and 321 backup plan
Short answer: money 😂
Well a No High Availability and 321 backup scenario leads to No money 😂😂😂