abbzer0

THANK YOU!!! You saved my bacon as well, back in business now! (I had to do it on both of the nvme drives for anyone else wondering, since there were in a pool)

I had overclocked successfully, but eventually had some random crashes I couldn't explain. Removed the XMP profile and ran memtest86 successfully, brought things back online, and stable since. Had issues with random docker crashes with iGPU transcoding, but a single 4k steam seems to do just fine with CPU transcoding instead.

I've been running into the exact same problems this past week. Last night I hard reset my unraid box again and completely disabled the docker service. Had an active SSH session into the box with a tail on the sylog file and I never saw anything showing the crash.. Last messages were at the 3AM'ish mark when it was moving something from the NVME cache to the array (also I've had crashes during the day, not just at night when that service starts.) I did, however, have the modprobe i915 piece in my go file. I am not using the iGPU acceleration for my Plex docker container at this time because of instability issues on the Alder Lake LGA1700 series board, so maybe that initial artifact when I had added it is causing the issues. I also re-seated my memory just in case.

As a follow up, I had issues rebooting and the USB functioning properly. Had to disable fast boot, and problem solved booting from USB upon reboot.

Okay, I researched multiple different threads for hours, and found a few things to get my new system working and thought I would share. GET UNRAID TO BOOT FROM USB -- My Motherboard saw the UEFI bootable thumb drive, but I couldn't get it to boot for the life of me. I used the Unraid imaging utility and tried making a new thumb drive (I was continuing to use my existing one, as this was an in-place upgrade), and couldn't get it to boot. -- The FIX? Rename the "EFI-" directory on the flash drive to just plain "EFI" - problem solved. - Also, you can't enable legacy compatibility mode (CSM) on this motherboard if you are using the iGPU - I guess something about it prevents you. Essentially you need a GPU card if you want to support legacy compatibility mode. See ASUS article UPDATE OS TO RELEASE CANDIDATE - I read some other threads that mentioned upgrading to 6.10.0-rc2. I'm not sure if I NEEDED to do that or not, but I did. And for those of you who don't know how to upgrade from the stable release train to the Release Candidates (RC), you go to TOOLS --> UPDATE OS --> SELECT "NEXT" ON BRANCH . NOTE: Be sure you backup your boot/thumb drive FIRST before upgrading! (Go to the "MAIN" menu, click on your boot device (Flash) and then click the "FLASH BACKUP" button. It will download to your PC.) GETTING iGPU TO WORK FOR PLEX - I found this thread helpful to get this piece working. I was originally unable to get anything to be created for /dev/dri . I did the following to get my i915 graphics to show up, and have the directory get created. -- Blacklist the i915 drivers so they don't load by unraid: "echo "blacklist i915" > /boot/config/modprobe.d/i915.conf " -- I then installed the "Intel GPU TOP" plugin. -- FYI - After doing the two above tasks, I now have the /dev/dri directory, and inside it are two files "renderD128" , "card0" , and a directory "by -path". Both of these two things seem to help get the iGPU recognized and working properly. GET PLEX TO USE HARDWARE - I then edited the docker container (advanced) for Plex and added the text "--device=/dev/dri" to the "Extra Parameters:" field. I hope this may help anyone who is looking to upgrade to the latest Intel Alder Lake CPU. So far, so good, this newly upgraded rig is FAST... I did have to re-enable FTP after the upgrade. I also installed 2 NVME drives to use as a new cache, but am not sure exactly how to migrate everything from the old cache pool to the new one. Part of me wonders if I just move everything over, and then edit the docker containers to reference the new path, so if anybody knows, please feel free to chime in here. Thanks!

Agree on the refresh stuff. It may have simply been the version I installed, but the "HowTo Guide" on disabling that was either dated, or didn't show up in mine. However, I made sure that I had no other machines/iPhones, etc.. still logged into the WebUI. I wouldn't expect that the auto-refresh would slow down to THAT much of a crawl. Also, even during normal operations (no parity checks running) I felt the UI was very sluggish. I would also eventually have a lot of memory usage over time, everything had to start buffering. Lots of strangeness afoot. And I think version 6 seems to be faster. I just found a screenshot of running parity checks on V5, and it was hovering around 63MB/s. (Granted mix/match of onboard/LSI connections, BUT..) I can't believe how quick the 120+MB/s checks are on Version 6!!

Okay, I felt the need to post this to the UnRaid forums to help benefit others. I never found this SPECIFIC example anywhere I searched. This is the result of a few weeks spent troubleshooting performance issues. My original problem statement was that all the sudden, while I had historically been getting fast parity check speeds of 50-60MB/s, it went down to 12-15MB/s. VERY slow speeds. There were a few things in play at the same time as I had recently upgraded from UnRaid 4.7 to version 5, replaced an aging 500GB drive with another 2TB drive, and installed the Dynamix WebUI (which looked very nice!!) and a few Dynamix plugins. Here is my hardware configuration: GIGABYTE GA-H81M-H Motherboard, 4GB Ram Processor: Intel Celeron - G3220 Haswell Case: Norco 4220 LSI 9211-8i Raid Controller (LSISAS2008 Driver) FWVersion(17.00.01.00) , ChipRevision(0x03), BiosVersion(07.33.00.00) - operating in IT mode Drives: 6 2TB Hitachi Deskstar 5300k Coolspin Drives, 1 Western Digital Red 2TB , 1 2TB Samsung HD204UI Drive A few other things I found out along the way: My original 2TB Hitachi drives are not advanced format, and use the 512k block, starting at sector 63. ( Deskstar 5K3000 HDS5C3020ALA632 ) The Samsung HD204UI drive is fine, but there is a firmware update that should be applied, due to SMART "stuff" potentially corrupting data. (I never had that problem, but found out along my troubleshooting.) The 2TB Western Digital Red (WDC_WD20EFRX) is another story!! Since it was not over the 2TB limit, I never selected the advanced format 4k alignment. Well, there I would be wrong!! This drive is of the advanced format type. While I determined this via a few ways, I believe the best way to determine the alignment capabilities is through the Linux command line. (Note, the first time I pre-cleared it with the preclear_disk.sh file, I used the 512 bytes alignment, and should have used the 4k, which is why you still see it starting at sector 63, and not 64. I am RE-preclearing it right now.) If you perform an "fdisk -l" you see all your drives there. You will notice on the Sector size, this one shows the logical/physical. All my "normal" drives without 4k alignment show 512 bytes / 512 bytes. However, the advanced format 4k alignment one shows 512 bytes / 4096 bytes. This helped clue me in on what I was missing: Disk /dev/sdb: 2000.4 GB, 2000398934016 bytes 1 heads, 63 sectors/track, 62016336 cylinders, total 3907029168 sectors Units = sectors of 1 * 512 = 512 bytes Sector size (logical/physical): 512 bytes / 4096 bytes I/O size (minimum/optimal): 4096 bytes / 4096 bytes Disk identifier: 0x00000000 Device Boot Start End Blocks Id System /dev/sdb1 63 3907029167 1953514552+ 83 Linux Partition 1 does not start on physical sector boundary. I completely removed that drive from the system and performed a rebuild with a non advanced format drive, but still had slow rebuild speeds. I tweaked all sorts of things in the BIOS to no avail. I have 4 slots in my Norco 4220 running from a reverse breakout cable directly from my onboard sata connectors (Amazon part " Discrete SATA to SFF-8087 Mini SAS Reverse breakout cable " ). I mixed and matched drives onboard, and connected to the other slots that connect to the LSI card, trying to figure out what was going on. Well, I'll cut more to the chase here. I found some posts that mentioned the Simplefeatures plugin (which I am NOT using) ended up causing slow parity checks. I found one post where someone had slow parity checks using Dynamix and plugins, but it never stated whether removing it or not resolved their issue. I suspected that it was possible my Dynamix upgrade, or the installation of a plugin coincided with my slow speeds, although I didn't have anything in the logs to suggest that. Here's where I can't show 100% the root cause, but it was one of a few things. I copied my config files from my USB thumb drive and wiped it, performing a clean install. Instead of going directly back to the current release of Unraid 5, I decided to install the Unraid 6 beta 6. I thought "why the heck not" and also read that it is using a more updated Linux Kernel. At present I have zero drives plugged into the onboard SATA ports, but have all drives connected to the LSI controller. Whether it is the new(er) version of Unraid, or just the removal of the Dynamix webUI and plugins, I can't say for certain. What I CAN say, is that as soon as I completed my fresh re-install of the Unraid OS (100% Virgin with no addins, etc..) my parity check speeds were at 123MB/s!!! That is AMAZING!! I don't even recall them going that fast on my older version, but at the time I had mixed/matched drives between the LSI card, and the onboard ports. Once I finish pre-clearing my WD Red drive again, I'm going to run it from the onboard slots, and see if there is any difference in performance. If there is, I will probably replace the motherboard (albeit only a couple months old) , and get another LSI card since I have been so extremely happy with the performance!! I hope all the information in this thread might be able to help someone else. I know it is slightly long-winded, but I hoped to get as much detail (model numbers, etc, etc..) into it, so hopefully could come up on relevant searches in the future. And now, it's time for me to start rocking my Unraid server again with great performance!!!!