lovaan

Along with your 2.5gbps nics and switch don't forget to get some cat6a patch cables. Technically you could get up to 10Gbps through cat5e if the cables are high enough quality and the runs are short enough, but spending a few extra bucks just to make sure makes sense to me.

Theoretically it should be 10gbps.

I've some issues after my last reboot where the system isn't finding eth0. From what I can see it appears to be loading the driver. I deleted my network.cfg and let it recreate it just in case there was something funky going on. It's a dual nic system so I tried using just the 1g port and then just the 2.5g port with no success. Any guidance would be much appreciated. tower-diagnostics-20250124-1521.zip

Looking at the spec sheet for the 9400 it looks like it provides 1 lane for each sata connection. I think the max data transfer for a sata3 connection is 600MB/s and the max for a pcie3 lane is 1GB/s

Assuming there's no important stuff like appdata or system shares on it I think you can just stop the array, click on the cache and remove pool.

You might be right with what you found, but I just read this stack exchange question from 2020 that might pertain to the usage too. https://superuser.com/questions/1582966/difference-between-filesystem-btrfs-size-with-df-btrfs-and-baobab

I was curious myself about this. I've never much paid attention to the overhead of a filesystem. So I talked it over with my AI and it comes down to this: The 1GB mentioned is the allocated space for metadata. However, due to how BTRFS RAID1 works, it generates and stores multiple copies of metadata across drives, which can lead to higher actual usage. This includes not just primary metadata but also snapshots, checksums, and various system overheads needed for data integrity. With the metadata ratio set to 3, each piece of metadata is stored three times across your drives. So with each drive storing its own metadata along with the metadata of the other 2 drives, it adds up.

Just so satisfy my curiosity, when you swapped sata cables how did you do it. Did you just unplug the sata cable from the hdd and plug it into the ssd or did you swap it on the motherboard too? If you swapped the cables at both the drives and motherboard there's a slim chance maybe you have a bad sata port on the motherboard. If you just swapped them at the drives then I'd assume your ssd is failing regardless of what SMART says. Either way if I was having this issue I'd backup my cache drive to the array if possible.

JMB585 only natively supports 5 drives and uses a port multiplier/replication to get it up to the 8 drives it advertises. This could lead to some headaches. I think the ASM1166 has a firmware that fixes ASPM.

If you look closely at the Suggested Fix section of Fix Common Problems it's telling you that since you have your appdata share set to Cache -> Array and not just have cache as the primary with no secondary, when mover runs applications will have a performance issues. You could ignore it if you don't care about application performance while mover is running, or just set the appdata share to cache as the primary and not have a secondary to remove the error. Also I think you'd have to manually move the appdata files from the array to the cache drive or you'll get a warning that appdata files exist on the array, or if you're not confident doing it manually you could temporarily set the appdata share to Array -> Cache and run mover, then switch it back to cache only.

I'd wonder if your beelink has multiple usb controllers and moving it to another usb group, maybe the one on the front, would make a difference.

I'm not very familiar with the live memory tester, but if I were to use it I would start my array in maintenance mode before running it, or better yet run the memtest provided in the unraid boot menu on startup. If you do in fact have errors on your ram then it's more than likely a bad stick of ram or a little less likely a bad ram slot on your motherboard, or if you're lucky maybe you just need to re-seat your ram.

Everything worked as expected. Also thanks for the tip on using a mirrored special vdev. Knowing that I'll hold off on adding the special subpool until I can mirror it.

NAME STATE READ WRITE CKSUM storage ONLINE 0 0 0 raidz1-0 ONLINE 0 0 0 sdh1 ONLINE 0 0 0 sdg1 ONLINE 0 0 0 sdl1 ONLINE 0 0 0 raidz1-1 ONLINE 0 0 0 sdm1 ONLINE 0 0 0 sdk1 ONLINE 0 0 0 sdj1 ONLINE 0 0 0 raidz1-2 ONLINE 0 0 0 sdb1 ONLINE 0 0 0 sdd1 ONLINE 0 0 0 sdc1 ONLINE 0 0 0 cache nvme0n1p1 ONLINE 0 0 0

I've added a cache subpool to my zfs pool. I now realize I'd rather have a metadata subpool, but I don't see a way to easily remove the cache subpool or convert it to metadata. I'm I missing something or would I have to re-create the entire pool?