March 11, 20251 yr Recently, I ran into an issue where my Cache pool's SSDs are behaving strangely. I run Unraid on Dell R730xd and have two NVMe SSDs (one Samsung 980 and one WD Blue SN570) plugged into ASUS Hyper M. 2 x16 PCIe extension card. This was running for some time alright, but after a recent power down suddenly issues showed up. First the WD SSD was not seen by the Unraid, so I have removed it from the Cache pool as I thought it got fried, but then I tried to reseat it in the extension card and got to a point where the system cycles between two states: 1) WD SSD is seen by Unraid and can be mounted using the "Unassigned Disk Devices", but at the same time the current Cache pool's SSD Samsung 980 can not be mounted and gives "Unmountable: wrong or no file system" error. 2) WD SSD is not detected by Unraid, and then the Cache pool is working as expected and Samsung SSD is mounting all alright. I originally thought that it might be a problem with the PCIe extension card, so I ordered a new one. It is also a newer revision (original PCIe card was Gen2, and the new one is Gen4), but unfortunately, the issue persists. By any chance, somebody has some ideas about what could be wrong and how to fix it, please? Diagnostics attached. Currently, I am running with detached the WD SSD in order to get the Samsung one mount. tower.cera-diagnostics-20250311-1115.zip
March 11, 20251 yr Community Expert When both devices are being detected, post the output from btrfs fi show
March 11, 20251 yr Community Expert 2 hours ago, AuHau said: ASUS Hyper M. 2 x16 PCIe I have that. Does your BIOS let you split up that x16 slot?
March 12, 20251 yr Author @JorgeB here is an output: Label: none uuid: e7f59792-310d-4a30-8b36-ce710793a617 Total devices 1 FS bytes used 9.64GiB devid 1 size 20.00GiB used 17.52GiB path /dev/loop2 Label: none uuid: 38f13438-862b-433c-acae-66af620a2241 Total devices 1 FS bytes used 1.31MiB devid 1 size 1.00GiB used 126.38MiB path /dev/loop3 Label: none uuid: 0cc578c8-2d7e-4960-a10c-16ba8fa60acb Total devices 1 FS bytes used 344.80GiB devid 2 size 931.51GiB used 354.03GiB path /dev/nvme1n1p1 devid 3 size 931.51GiB used 440.06GiB path /dev/nvme0n1p1 @trurl yeah, I have setup 4x4x4x4 bifurcation. It worked before... I will try different PCIe port.
March 12, 20251 yr Author A different PCIe port did not solve it. I had to use the old PCIe NVMe expansion board though, as the new gen is just too big to fit into the other PCIe port 😳 This time, it is ignoring the WD SSD, so I am attaching new diagnostics. So far this "disappearing" of the WD SSD happened only on the old expansion board, so maybe this problem is related only to the old expansion board? tower.cera-diagnostics-20250312-1237.zip Edited March 12, 20251 yr by AuHau
March 12, 20251 yr Community Expert Solution Data Metadata System Id Path single DUP DUP Unallocated Total Slack -- -------------- --------- --------- -------- ----------- --------- ----- 3 /dev/nvme0n1p1 367.00GiB 10.00GiB 64.00MiB 554.45GiB 931.51GiB - -- -------------- --------- --------- -------- ----------- --------- ----- Total 367.00GiB 5.00GiB 32.00MiB 554.45GiB 931.51GiB 0.00B Used 344.27GiB 535.91MiB 80.00KiB The pool is only using one device, but the other one has the same UUID, if the pool contents are OK, you need to wipe the other device, you can then add it to this pool or create a new one with it.
March 12, 20251 yr Author Yey! That helped! I was thinking about this posibility, but was keeping it as last resort as I was worried about potentially loosing some data. Anyway - thanks a lot!
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.