rangusT

After a couple of Unraid updates I stopped experiencing this behavior. I attributed it to the updates but never found a conclusive answer.

I recently bought some new gear to add 10gbe to my unraid server and pc client. I have an Intel x520 sfp+ card on both sides and are connected to each other via an spf switch. I have an nvme drive in the server which I benchmarked using the DiskSpeed docker and got 2 gigabyte per second read and 2.7 write. I also have an nvme on my Windows PC so that should not be a bottleneck. I shared the drive using the unassigned devices plugin and found the max speed I can hit is about 250MBps uploading to the server and 400MBps downloading from the server. I was expecting speeds closer to the link max at 1GBps. I installed iperf3 on the server and get around 5 Gbps or 7 Gbps with the -R flag on 1 thread. If I up the threads to 3 or more iperf3 will saturate the connection. I'm not really sure where to start. There are a ton of threads on the forum with various advice surrounding 10gbe but nothing conclusive in improving the transfer speeds. morioh-diagnostics-20240107-2002.zip

I started experiencing a very bizarre issue recently. I woke up one morning to find that my entire network was down. Initially I assumed an ISP outage but found that I was still able to connect to my VPN (the VPN server runs from my router, not Unraid) from my phone via 5g. After much experimenting I discovered that once I unplugged the ethernet from my Unraid box my network instantly came back online. It has since happened two more times and pulling the ethernet brought my network back online both times. I want to say this started happening after upgrading to 6.12.0 but I am not completely confident in that timeline. I have since upgraded to the latest 6.12.3 and still experienced the issue. I can provide diagnostics but I'm not sure how useful they'll be. Once this issue occurs I am unable to access Unraid until I hard boot the server. I connected an hdmi to attempt to see the console but I can't get video until I hard boot the server. I normally run the server headless. I am unable to get network access to the server as well until a hard boot is performed. I'm hoping to find some advice on how to troubleshoot this issue. morioh-diagnostics-20230720-2008.zip

I started the rebuild as you suggested. Now, a different disk is throwing a significant number of errors. I attached diagnostics.

Unfortunately I didn't think to grab diagnostics before shutting down the system. In the main tab there were counts in the errors column on the affected drives. I want to say between 4-5000 on the parity2 and a few hundred on disk5. I'm assuming these errors triggered the disable. If I do a rebuild on top, which way would be safest? Parity first, or data drive first? Would it make sense to use one of the empty drives to rebuild into rather than on top of the existing drives?

I noticed today that two of my drives are marked as disabled. One of them is a parity drive, the other is a data drive in the array. I am running dual parity. I am looking for advice on properly assessing whether my drives are faulty and needing to be replaced. I ran short SMART tests on both drives and that came back normal. Also of note, I have two extra drives that are currently in the array but empty, so I could repurpose these to replace the two disabled drives if they are bad. If I need to do that, what order would be best to tackle this? Party or data drive first? I have not had a drive failure before in unraid, so I don't want to take any uneducated steps that could permanently lose data.

Are people storing their plots on storage that's part of the array or on unassigned devices?

Getting the same issue here as Fiservedpi.