JonathanM

How does it perform wired?

Motherboard groupings are controlled by the chipset. There are some options you can try to break up the groups, but they don't always work and can produce instability in rare cases. Play around with the various PCIe ACS override: options in the vm manager settings. If you find a setting that seems to work, read up on it to check for signs that it's causing other issues before you run that way long term. If you can't get it working with those options, a different motherboard may be the only solution.

The only good thing reiserfs has going for it is the extreme resiliency for recovery. If you change the file system type back to reiserfs, start the array in maintenance mode, then run the reiserfs check command reiserfsck --check /dev/md2 Note that there is a space between --check and /dev/md2 Hopefully it will give further instructions, I'd capture that output and post it here before proceeding further. This command may take a long time to analyze the drive since it was formatted with another filesystem, let it finish the check.

Why? Both plex and jellyfin have docker containers available which utilize the system resources much more efficiently than piping them through a VM.

The VM's will likely perform better with far fewer cores dedicated. Remove all except 1 pair, 7/15, and see how the VM feels. Add 6/14 and test again. Repeat adding from the high numbers down until the VM doesn't perform better, then back off one pair. Always leave 0/8 unassigned, the host needs it. If you have multiple VM's running concurrently, you may need to test different combos, but always try less cores rather than more. The more resources you can let the host use, the better it can serve the VM with I/O and other emulated services. Same goes for RAM, even more than CPU cores. RAM dedicated to the VM is lost to the host, so you never want to allocate more than is absolutely necessary. RAM available to the host will be used to cache I/O, which really helps the VM feel snappy.

The motherboard power distribution circuits play a role here, so some boards may handle it better than others, and marginal boards may be stable for a while but as they age may not be able to maintain stability. Not DDR5, but I had a board a few years ago that initially ran fine with all 4 slots, but after a while it started crashing and failed memtest. All 4 sticks individually tested good, and all slots individually tested good, but as soon as more than 2 slots were populated it failed memtest.

That is not a universal thing, granted it's most of the world, but there are areas where electric usage is not wasted because it's needed as heat anyway, or renewable production is excessive for the local demand. Your general sentiment is valid, but there are cases where the benefits of continuing to use outdated hardware to the fullest outweighs the waste of discarding it for newly manufactured product that needed resources and energy to produce. The reduce, reuse, recycle triangle applies.

I do NOT speak for Limetech (not Limewire, that's an old file sharing thing) but I have a question. Would you be OK with an online license verification option? The biggest reason I believe we are still locked in to flash GUID for licenses, is the cheap nature of a unique token that can be verified completely offline. Other options in that space add a minimum of $30 for the hardware, and it would have to be sourced and maintained by Limetech, which adds a bunch of cost and complexity to the licensing scheme, not to mention the extreme PITA when a dongle fails, is lost, or broken. Currently if the flash drive fails it's a relatively low cost user replaceable item, and the key replacement is handled by the automated system. Limetech even maintains a cloud backup option to make it even easier. I don't see a whole lot of good options for people who want a totally firewalled system that runs without contacting Limetech on startup. What did you have in mind? @SpencerJ, care to chime in here?

You MUST load test to find that out. Batteries can (and often do) fail in a state where voltage is normal during charging, but as soon as a load is applied, the battery falls over almost immediately.

Shouldn't take weeks if it crashes before 2 days. Enable half of your normally running containers. If it crashes, divide those in half. If it doesn't, disable that half and enable the other set. I recommend printing a list and noting the start time of each container and notate crash times, keeping track of which containers were running at that point. Shouldn't take more than a few cycles to narrow it down, unless it's a combination of containers that only crash when they interact, or you have 100's of containers. Bonus is, you get to continue using critical containers.

Physical slots don't matter, logical slots in the Unraid Main GUI do matter.

If you can manage to simultaneously feed your Unraid uninterrupted power, then when the UPS goes into backup mode Unraid should start the shutdown process. If you connect something with a similar electrical load to the UPS, maybe a small space heater, then you can see if the Unraid shutdown beats the UPS running out of power. It's entirely possible Unraid's shutdown process is timing out before something can get stopped, forcing Unraid to shut down uncleanly. If you are running VM's, they may not be stopping properly, or take too much time. Personally I run apcupsd software inside my guest VM's in slave mode, so when the host detects a power outage the VM's immediately start the shutdown process.

Your current layout of 2 spinners and one SSD is typical. If you are only using it for bulk storage, you don't need SSD, but if you want to run containers and VM's, it's pretty much a necessity unless you are prepared for a severe drop in performance.

Docker containers and VM's will be extremely slow if you try to run them from the parity array. Is there no way for you to add a SSD of some flavor? Empty PCIE slot? M2 slot?

Yes, post a reply in that thread asking your question.

The easiest way would be to temporarily put all the old drives in the new system, they don't need to be in the bays as long as you can run SATA and power to them. How are your current drives assigned? Array, pool? How are you planning to assign the 2 permanent drives?

Use -L

Many BIOS's have three states for power on. Always on, Always off, previous state.

I think that would need something running inside the same LAN to send the signal. Perhaps if there is a local path detected it could offer the option? I don't see a way to accomplish this if the server and and connecting machine are on different LANs.

Normally you would move all the drives from the old server to the new. Please describe your setups in more detail.

Shared, yes, but also remember the VM "motherboard" is running in code on the host. Restrict the motherboard, hamstring the VM.

I'm guessing they want the backup for some containers sent to a different folder so that the backup can be archived elsewhere. I can also see a use case based on importance of the specific container. Critical backups sent to redundant local fast storage, "nice to have" backups sent to cheap low priority bulk non-redundant destinations. Both desires can be handled manually with scripting, but it would be easier if the backup destination could be made more granular. Backups for your backups.

https://forums.unraid.net/forum/75-multi-lang-forums/ https://forums.unraid.net/forum/78-spanish-español/

This thread will get lost in the noise in a matter of days. If you really care to help others, post this in the support thread for the container in question, that way your post will be read by others seeking support for that container.

Try a different USB that you are positive hasn't been used. GUID's are tracked.