I keep config backups of the config as well but my boot drive is a pair of mirrored Intel D3-S4510’s that are underprovioned by 50GB.
Rollback time will technically depend on the size of the change since a rollback only replaces changed files so that will depend on how far back you are rolling back to. My largest snapshot references about 2GB and is a little over 4 months old, I would estimate about 5 minutes at most rolling back to that point.
I’ve never been comfortable virtualizing a router. Nearly anything else is not an issue.
On OpnSense no. I previously did not have snapshots created. I have on FreeNAS when I had a setting issue that came up. Fundamentally it should be no different.
I have had to reinstall OpnSense before and would rather not go through that again if at all possible.
No clue. Don’t use it myself. I am able to access the website so I would assume the rest is up and working.
Your ISP may be blocking access (many ISP’s do not allow hosting things on home connections so check your TOS.) Why they would block access to the website I have no clue but they could see allowing it as allowing people to host a service.
Disconnect your phone from wifi and see if you can access DuckDNS. You can also do some troubleshooting and testing https://www.golinuxcloud.com/how-to-check-if-isp-is-blocking-website/
Also depending on your router if you can setup DNS over TLS or DNS over HTTPS and use two different DNS providers not controlled by your ISP.
If you’re using an ISP provided Modem/Router you should probably consider setting up a proper firewall router and get just a modem or put the modem/router in bridge mode so you don’t double NAT.
A snapshot is an image of the complete system at the time of creation that tracks changes and can be restored if needed.
https://www.howtoforge.com/tutorial/how-to-use-snapshots-clones-and-replication-in-zfs-on-linux
Software RAID, yes. Hardware RAID, NOOOOOOOO!!!
Without knowing what the hardware is in the system I need to stress that before getting started.
Software RAID is controller independent, if the MoBo dies tonight you can get a replacement of whatever that is cheap or even grab a HBA (Host Bus Adapter) and hook the drives up. Hardware RAID requires an identical controller with the same version of firmware in most cases. In a home environment hardware RAID is a major pain, most corporations have spare hardware on hand if there is a failure so it’s not as much of an issue. Also JBOD (Just a Bunch Of Disks) is not the same as a HBA so using a hardware RAID controller with that turned on for a software RAID is asking for issues. For software RAID in OMV you will be looking at something like BTRFS https://en.wikipedia.org/wiki/Btrfs or using a plugin to add ZFS https://en.wikipedia.org/wiki/ZFS
I come from a FreeNAS/TrueNAS background using ZFS and SAS2008/SAS3008 HBA’s.
There are different version of RAID array’s, you need to look them up and decide what works best for your data. What works for me or anyone else is irrelevant. No RAID is not a substitute for good backup policies, but it saves a potential TON of work doing a restore and the headaches associated with it. https://en.wikipedia.org/wiki/Standard_RAID_levels
RAID 1 or Mirrors is great for data integrity. You could put all 7 disks in a RAID 0 or mirror configuration and it would take 7 disk failures to lose your data.
RAID 0 or Stripes is great for speed and storage capacity, the data is split into chunks and with all 7 disks you would have maximum capacity of the total sum of all the disks, one disk fails and all your data is gone like a fart in the wind.
RAID 0+1 or Striped Mirrors combines the two above concepts. In some ways you get two disk failure and in others you don’t. Take 4 decks of cards, set two side by side, those are your mirrors. Take the other two and put them above the other pair, those are the stripes. Lose two mirrors and the data is gone, lose a single mirror in each stripe and the data is safe. You can mirror and stripe to your hearts content, want 3 mirrors with two stripes it takes 6 drives.
RAID 5 or RaidZ 1 gives you one disk failure redundancy, how it functions depends on the raid controller or software utilized but basically data is split in to chunks and you have the ability to replace a single disk and fully recover. If a second disk fails your data is gone again. Some people consider this dead as the odds of a large drive having an issue during a rebuild is higher. https://www.servnetuk.com/insights/ure-is-raid-5-dead
RAID 6 or RaidZ 2 adds a second disk failure redundancy. If you want maximum space this is what you should probably use with failure redundancy. It’s not super fast but unless you need the speed for anything in particular it’s a good trade off. With 7 drives this is a good option.
RaidZ 3 Adds a third disk failure redundancy. This is what I use for most of my pools. When rebuilding you are stressing the disks to the maximum and that is when you will find a second issue pop up. In a RaidZ1 your data is gone, in a RaidZ2 you lost your last spare. In a RaidZ3 you still have one failure that will save your bacon.
I have multiple pools. I have the majority of my regular data on a RaidZ3. I have a couple striped mirrors for things like VM’s, Docker containers, and things I want a little more speed on.
You can also do things like RAID 0+5 or 1+5. In ZFS this would be called RaidZ1 with two vDev’s.
I use TrueNAS because of ZFS, BTRFS is still a work in progress in comparison.
I imagine about now you just about had your head explode, it’s tough to make this decision because it puts you down a path. I started when BTRFS was an infant, I stick with ZFS/TrueNAS because I know how things work and it suits my needs. You need to decide what will work for you and you may find that OMV isn’t the best path going forward. Take the time to really make a decision about what your needs are today and what they will be in the future 5 to 10 years plus down the road. Don’t just say, well (random person on the internet) says this is best so it’s what I will do. Every decision is about tradeoffs and compromises.
Once you make a decision one of the best things you can do is do a burn in on the drives. I do it with BadBlocks and run smart tests. I just got done with a batch of 8TB SAS drives and found 4 with issues that were replaced before I even put them in to production. If a drive survives a torture test early on it’s less likely to have issues later on. I upgraded from a pool of 4TB drives and out of 11 drives I had 1 failure. When you get to the point where you are thinking about expanding your pool with new drives you will want to buy from a couple different sellers, at different times a week or two apart, or different manufacturer’s.
One other thing I will add… Toshiba is verboten in my world. They don’t replace, they give you a check and you have to buy at current prices and it’s more of a PITA than I want to deal with. BackBlaze is also a great place for information about drive reliability. https://www.backblaze.com/cloud-storage/resources/hard-drive-test-data
No problem.
Be sure to set a cron job for updating the DNS Blocklists, running scrubs, smart tests, and trim if you are using a ssd.
I am working on a way to do snapshots for the system as well for super simple restores that are automatically created. Before an update you WILL want to have a snapshot available. Once snapshots are automated I will probably start doing automatic updates. It basically creates a system similar to how an immutable OS works, if something goofs you can simply roll back to a previous working state and pause updates. I previously had a script I setup via a cron job but the system removed the job during an update.
Try setting a tunable of:
net.link.ether.inet.max_age = 540
By default FreeBSD sets a 20 minute ARP maximum age while many ONT’s set a 9 minute maximum age. This will show up as performance issues initially then cause other issues.
Also Realtek NIC’s are notorious for issues with FreeBSD, so if you are using one replacing it with an Intel one is highly recommended.
Absolutely. Remember to enable a cron job in the OpnSense - System > Settings > cron to update the Blocklists. Otherwise you are likely out of date.
Which one? Lol.
TrueNAS is mostly via the webgui but I use SSH when needed.
I have two Ubuntu 26.04LTS VM’s running on TrueNAS. One is going to be my web host the other is for FrigateNVR and local AI processing. They are both managed via SSH. The web host has a FIDO2 ssh key for access, the other is a standard key and is not web accessible. Both can be accessed via remote desktop when at home.
Docker containers are managed via DockHand. I have about 50 containers running across the above hosts right now.
HomeAssistant is managed via the webgui and SSH when needed. It runs on its own hardware (Odroid N2+)
My traveling local host/ap/dhcp server is primarily managed over USB via SSH. Files are managed via SFTP.
On my wall mounted stuff at home I use HomeAssistant and turn the charger on when the batteries get to 50 percent and off at 80 percent. Not really an option for this device while on the go though.
Yes on the solar panels. No on the HOA. Small town life is good. I installed the system controller 3, IQ Combiner 5, and two IQ 5P battery’s already. Never pulled a single permit, nor had anyone tell me I couldn’t do it. 7.2KW with IQ8A’s will be on the roof soon.
Damn swype… I replaced the roof 2 years ago. And I mean that literally tore it off completely all (4 layers) and put a new one on by myself.
Yeah, I don’t trust any of the device batteries truthfully. This one has a little swelling because my stepson used it for a while and had the bright idea to leave it constantly plugged in. I will not be flying with it for sure and when I do travel it will not get packed where I can’t access it easily.
If the bootloader can be unlocked you can do a ton of things like this. If not you are limited to using Termux and the packages you can install without roof permissions.
You can still get the Google Coral (both the USB dongle and M.2/PCIe versions), but Google has effectively abandoned the official drivers. You’ll need to use community-maintained DKMS drivers if you go with the PCIe version on newer Linux kernels.
If you want a modern alternative, look into the Hailo-8 or Hailo-8L M.2 AI accelerators. They plug right into an open M.2 slot and natively integrate into recent versions of Frigate. Keep in mind that these accelerators ONLY process object detection inference - they do not handle video decoding.
For video decoding, your OptiPlex 7070’s Intel CPU has Quick Sync Video (QSV). You can pass /dev/dri through to your container, and Intel iGPU hardware acceleration will handle decoding the RTSP streams with minimal CPU overhead.
If you prefer an all-in-one approach, an Intel or Nvidia GPU can handle both video decoding and object detection (via OpenVINO or TensorRT), but a Micro PC limits your physical expansion options.
Storing recordings on a TrueNAS share over NFS or SMB works just fine. Mount the NFS share directly in Proxmox or the VM host, then pass the directory through to your Docker container.
Never used VyOS so can’t help there. I do use OpnSense, TP-Link Omada EAP-650’s (with an isolated vLAN for the guest network) multiple vLAN’s for cameras, iot, management, Trusted devices, and DMZ, along with Wireguard for remote access and since my ISP only gives me an IPv4 address I use a Wireguard tunnel to Route64 for IPv6 connectivity, a cellular connection for backup internet connectivity, CrowdSec, Intrusion Detection, Caddy, and UnboundDNS.
I used multiple different Router OS’ since around 2005 and settled on OpnSense years ago. I stick around because there is rarely an issue and the reporting system makes it easy to visually spot issues.
ZFS rollbacks only have to roll back the changed files so it’s a pretty quick process unless there were some major changes.
Yeah, that was the goal with the D3’s. I’m trying to get things to the point where if I’m no longer able to handle things that the router will stay updated and be simple to repair for my family.