January 6, 20233 yr I have cache pool of 5 SSDs for VMs. I was trying to set up iGPU passthrough and ran into some errors. Here is what happened: set graphics card to iGPU in VM template went to Intel GVT-g plugin and assigned a mode to VM tried to start VM but just kept trying to start, wasnt seeing any relevant info in logs tried to reboot server but nothing was happening - probably waited 20 minutes or so did a hard shutdown as I wasnt getting anywhere when server came back up, a parity check started but I cancelled it and then I saw the invalid pool config error since the pool was unmountable, i couldnt edit the VM config to set the graphics card back to VNC so I turned off VM Manager, rebooted again, but did not fix the issue. Ive also removed the assigned VM from Intel GVT-g plugin Diagnostics are attached. I do have a VM backup (untested though as I just took it today) but I would prefer to try and fix the issue instead of reformatting if possible. Appreciate any advice on next steps. Thank you for reading. Edit: Was searching around and tried the following - I stopped the array, unassigned the drives, started the array, stopped, and then re-added the drives and started again. Now I have a different error "Unmountable: No pool uuid". I rebooted to see if this would fix at all but it did not. omnius-diagnostics-20230105-1751.zip Edited January 6, 20233 yr by SOULV1CE Troubleshooting
January 6, 20233 yr Community Expert You appear to have the two connections going to the same enclosure, if it's a dual expander or dual controller enclosure connect just one cable and post new diags.
January 6, 20233 yr Author Yes I have dual channel transmission cables from my backplane to HBA. I shutdown, unplugged one of the cables, rebooted and now magically the VM cache pool is valid again! So it seems it was something with the dual channel that was causing the issue maybe? Is it ok to plug in the second cable again? New diags are attached. omnius-diagnostics-20230106-0926.zip
January 6, 20233 yr Community Expert 4 minutes ago, SOULV1CE said: So it seems it was something with the dual channel that was causing the issue maybe? You were not using dual link, dual link is two connections from one HBA to the same expander on a backplane, in this case the HBA was connected to two different expanders on the same backplane (or an enclosure with dual controllers), that's not supported by Unraid since the devices will be duplicated.
January 6, 20233 yr Community Expert That type of connection if for redundancy, not speed, dual link on the other hand improves available bandwidth.
January 6, 20233 yr Author 14 minutes ago, JorgeB said: That type of connection if for redundancy, not speed, dual link on the other hand improves available bandwidth. Ok thank you for the explanation! I have a BPN-SAS3-846EL1 backplane using both the J49 and J50 connections to a single LSI 9300-16i HBA. I unplugged ports 2 & 3 on the HBA. So just for me to be 100% clear, there is no redundancy benefit for me to plug these back in because this is not supported in Unraid?
January 6, 20233 yr Community Expert Solution Ahh, so you had the 4 ports connected to the expander? If so, yes, that's not a valid config, HBA to SAS supports single or dual link, not quad link. Since that HBA is basically two 9300-8i on the same board you had for all purposes 2 HBAs connected to the same expander, Unraid won't support that.
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.