Rhuarc

I think I understand, but are you saying basically to stop the array, remove the second device, reimport the pool, then... actually I'm not exactly sure what I am supposed to do. I don't want to screw this up. I apologize for the confusion! I tried just removing the second device and starting the array and it says too many missing or unassigned devices. I am guessing maybe I have to drop teh number of devices on the pool back down to 1 for now?

This is what I got when I assigned the 1TB that had been in the pool to the second slot in the pool. It will not let me start the Array with the second (1TB) device in slot 2 of the pool.

Here are the results of the scrub with only the 480GB drive, no errors found. Also grabbed an update Diagnostics incase that helps. Edit: I did just get this error in the log. No idea if this applies or is something completely different. milkyway-diagnostics-20250515-0021.zip

Is this going to delete the stuff on the pool, or should I be fine? Theoretically both devices should have all of the files from the pool, correct?

Thank you for helping out!

I am attempting to finish upgrading 2 cache pool drives. I had dual NVMA M.2 480GB ADATA drives in RAID1 for my cache pool. I was running out of space and wanted to upgrade them both to 1TB drives. I started following these directions from the FAQ thread: The first drive seemed to upgrade fine, but every time I attempt to replicate the instructions for upgrading the second drive it gives me errors about a missing disk. I have attached Diagnostics from the current status which is with 1 drive upgraded to 1TB, and the 2nd drive still being the 480GB. I am hoping someone can take a look and let me know if it seems to be in the correct state for me to replicate the instructions to replace the second drive. Once I attempt it I can take additional diagnostics if the operation fails. Thank you! milkyway-diagnostics-20250514-1251.zip

I recently upgraded the GPU from a 1660Ti that was working great for transcoding in plex, to a 1080Ti. I shutdown the server, swapped GPU, and now my plex docker has disappeared and the GPU is not showing in the Nvidia Driver plugin, and the GPU Dashboard isn't showing it either. It IS showing in the System Devices, and I can select it for the GPU stats dashboard but the dashboard shows this error: I am also getting these errors trying to shutdown the server to pull diagnostics. Diagnostics attached. Hopefully they can help shed some light. I put back in my old 1660Ti and everything is working as it should. milkyway-diagnostics-20250323-1900.zip

Anyone have any ideas on this? Thank you!

I had been running the Backblaze Personal Docker container for awhile and it was working great. Something changed and it will no longer start with the following error. Not sure what is going on. I have deleted the container and image and reinstalled with the same result. Not sure what else I can do at this point so I am here to look for help from the experts. Thank you! Rhuarc

Thanks for the advice! Here is the one I am specifically looking at: https://www.amazon.com/Intel-RAID-Expander-Card-RES2SV240/dp/B0042NLUVE In this case I could connect 4 drives to my HBA, the HBA to this expander, and then my other 8 drives to this expander. I believe this can also be powered over a molex connecter. I would probably test to see if using power from an x1 slot would drop the speed of the x16 slot. If it does then I would kapton tape the pins and power over molex. If it does not then I would just power using the x1 slot. Oh, also, on these expanders do they use a different type of cable to connect the drives than what I currently have connecting the drives to the HBA? I know I still the the cable to connect between the HBA and the expander also.

Alright, so my idea for putting the graphics card in the second x16 slot and running it all the way down at x2 speeds and putting the HBA in the top 16x slot does not seem to have worked. The GPU, while recognized and seeming to be passing through correctly to Plex and doing transcoding, is not working correctly. The transcode seems to be software only even though GPU Stats is saying it is using the GPU. To test I swapped the 2 cards again and right back to full transcoding speeds. Basically now I need to figure out how to run all 12 of my drives only using the HBA. This will free up the second M2 on my second NVME drive in, and then the NIC can go in the PCIEx1 slot currently occupied by the M2 to NVME adapter. My questions are: 1) If I get an Expander, can I cover the pins in kapton tape and then put it into an X1 slot only to hold it and then power it off of a Molex? 2) Would I even need to do #1, or would the Expander just get power and the board would not even recognize that it has a card running therefor dropping the speed of the second x16 slot? 3) Would I then run 8 of the drives off the Expander and the other 4 off the breakout cable of the HBA, or if you have an Expander do you never have any drives connected directly to the HBA? Thanks again!

Thanks for this! I actually just did the switch. Now have dual 2.5GbE, the GPU is now showing 2 lanes instead of the 16 before I made the swap and added the NIC. As long as the 2 lanes are plenty for just transcoding in PLEX this should not be a problem! And as a bonus after running Diskspeed on the controller I actually gained around 10 percent performance on all disk reads and the Controller is no longer being the bottleneck! Before: After:

Yeah, this is my main problem! I do want the faster NIC. I do photo and video editing and keep my main library and files on the NAS. Being able to edit with little delay to the full size files is greatly increased with the faster direct connection. I think I have space to add the NIC to the bottom slot even with the 1060 in the bottom 16x slot. The main thing I need to check also is that the bottom 16x2 slot can output over a video card. I'm assuming it can and that the system won't think the HBA is a video device and try to output over that. I wonder also if the HBA will run faster in the top 16x slot that in the bottom one even when the bottom one is running at full speed. Is there an easy way to do a speed test on the HBA now and then rerun later? A parity check I know can indicate but I get wildly difference speeds during parity checks depending on what else I am doing on the server. I guess I could disable docker and vm's and run a 20 minute parity check to see what speed I get, then redo everything and run the same 20 minute parity check to see what speeds I get then.

Would it be any better if I swapped the 1060 into the bottom slot and put the HBA in the top slot? The onyl thing I use the 1060 for is outputting the GUI to a local monitor and pass through to a PLEX docker for transcoding. Theoretically this would probably use less bandwidth than the HBA? Thank you for your help!

Hello everyone. I am trying to figure out how to add a dual NIC 2.5GbE to my build. I have a Asus Prime B550-PLUS motherboard. an LSI SAS2008 HBA running 8 spinning drives. 4 spinning drives and 2 SATA SSD connected directly to the MB. I have 2 NVME, one on the MB and one using a PCIE to M2 adapter. I have put together a kind of diagram, and I am wondering if what I am proposing will significantly limit the speed of the HBA. The slot the HBA is in is normally a x4 slot (I believe this refers to speed) that drops to x2 if either or both of the bottom to x1 slots are occupied. I think I need 1 of these slots to accomplish all of the hardware i need. So the question is, will dropping the HBA to x2 instead of x4 slow anything down in a significant way? If there is anything I can clarify to help me get this optimized please let me know. Thank you!

Ok, so I just had another crash. Luckily the rebuild was completed so I didn't have to worry about restarting that. Here are the latest diagnostics. I am hoping it will show the same issue with the network so that I can start isolating it to that. I will switch to only the internal NIC and see if I get any more of a solid run time. Any good reccomendatiosn for an internal unraid NIC that is 2.5gbe? Thanks again! milkyway-diagnostics-20240718-1608.zip

I am, and I am also using a kvm that has a USB component that plugs into USB. Is there any chance that caused the issue? My problem with the nic is that I don't have enough slots and lanes on the motherboard to go with an internal 2.5gbe nic. Should a USB nic be fine or do they historically cause these kind of problems? If I were to remove something to make space for an internal nic are there any 2.5gbe that come highly recommended? Thank you!

I am currently doing a rebuild on a larger 16TB drive from an older 8TB drive. Everything was going smoothly for around 50% then it became unresponsive to both webGUI and GUI directly connected. I could still ping the server but SSH into it was not working either. Attached are diagnostic logs from right after the reboot. Thanks everyone! milkyway-diagnostics-20240716-0233.zip

Been running it now for a week and everything seems to be going great! I can get a couple extra of the sd cards to keep on hand in case one dies and no worries about needing to change keys!

I did a few tests writing to an sd card constantly for 30 minutes and inside the enclosure never hit above 50C. Obviously not a fully scientific test, but encouraging at least. I would be surprised if modern hardware like a card reader can't be good to at least 60C for extended periods.

I found one place that says it is rated to 60C, but I have no idea if that is reliable since it isn't listed on any spec sheet. I have a second reader and card coming this weekend. I will do a test using my enclosure do a sustained write sequence to the card to reay heat it up and stress test it and report back the temps it hits inside the enclosure right against the reader.

Hmm, Hadn't thought too much about it. I can do a test using a temp probe. I'll put a small hole just big enough for the probe and see what temps I end up getting. The Sandisk industrial card is rated up to 85C and 190TB lifetime. I think it should be fine. Worst case I will design the enclosure with holes in the top and bottom so it has more of a bumper appearance but lets the actual top and bottom of the reader breathe. I'll report back once I get this data!

Just got this reader and card and it is working perfectly. I already tried switching between two cards by running the usb creator and then copying all the files from a previous card and it works. Very easy. https://www.amazon.com/gp/product/B07G5JV2B5?ref=ppx_pt2_dt_b_prod_image https://www.amazon.com/gp/product/B07CV344WJ/ref=ppx_yo_dt_b_asin_title_o00_s00?ie=UTF8&psc=1 I was a little worried about the card just being open to the back, so I cam eup with a 3d printed enclosure for the reader with the card. Basically turns it into a full usb drive that I can but the zip tie and swap the card out easily enough. I'm just using some rubber sstrips right now to secure it inside the enclosure but at some point I will use something like Sugru to perfectly mold the inside of the enclosure to the reader with card.

Is there any chance that these crashes are related to the KVM that I have hooked up? https://www.amazon.com/gp/product/B092Q2DVC2/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1 I believe it acts as a hub or something like that. It has a video cord plugged in to the graphics card and a USB plugged in for mouse and keyboard support.

In doing some research it doesn't seem like Intel has ever made a USB NIC. I want to keep it at or over 2.5GbE, but currently have no leftover pci-e lanes on my MB as far as I am aware. In looking at the logs does it seem pretty solid that the NIC is what is causing the no responsive web gui? Also, if the web gui becomes unresponsive due to the NIC, if I have my server set to boot into the GUI on the connected Mouse, Monitor, and Keyboard that should theoretically still work, correct? I really need to figure out these crashes so I can get some stability back in my server. Thanks again!