Corrupt_Liberty

Docker and VM Manager were disabled at the time. That's not really the issue though as I was able to recover all of that from backup. Everything is back up and appears to be running normally but this weird issue writing to the cache drive is a bit annoying.

Yes it was. There are some remaining files from the Cache pool upgrade. I originally followed an old tutorial that had me using the mover to clear the pool and then move it back after I recreated the pool. This failed miserably. I was able to recover my appdata from backup, delete the docker image and reinstall my apps without too much fuss.

Okay, sorry for the delay. I tried pulling diagnostics and it wouldn't work. I've been meaning to update (the previous update bricked it) so I went ahead and did that first. I re-tested a file transfer and the issue still remains. kris-server-diagnostics-20220716-1647.zip

I originally had 2 SSDs in RAID 0 for my Cache pool. When the drives went on sale I decided to get two more and change the pool to RAID 10. Ever since it has been exhibiting some strange write performance issues. It starts writing just fine at 110-120 MB/s (limited by my network speed). However, after a few GB are copied the speed drops to between 20-40 MB/s. This never happened while I was using RAID 0 and can't for the life of me understand why it would start behaving this way now. Is this an odd limitation of the btrfs file system?

Also just noticed that my WAN and Public IP addresses are different. Any idea how to fix this?

Apparently this is an issue with my Plex docker as well. I am really at a loss as to what could have happened here.

I just changed ISP to a fiber provider in my area (Metronet). The only network change was to unplug the cable modem from my routers WAN port and plug in the fiber modem. Everything else has been left unchanged including my forwarded port on the router for my Unraid server. Unfortunately, I am now no longer able to access my server remotely. When I load the My Servers page from my 5G tablet or phone it says Remote access is available but the site cannot be reached. The My Servers plugin also says it is unreachable and I need to forward port 443. This was done months ago and I've verified it is still active on the router. php /usr/local/emhttp/plugins/dynamix/include/UpdateDNS.php -v (Output is anonymized, use '-vv' to see full details) Unraid OS 6.9.2 with My Servers plugin version 2022.05.24.1752 ✅ Signed in to Unraid.net as Corrupt_Liberty Use SSL is yes ✅ Rebind protection is disabled for unraid.net Local Access url: https://hash.unraid.net ✅ hash.unraid.net resolves to 192.168.0.154 Remote Access url: https://www.hash.unraid.net:[redacted] ✅ www.hash.unraid.net resolves to [redacted] Request: { "keyfile": "[redacted]", "plgversion": "2022.05.24.1752", "internalip": "192.168.0.154", "internalhostname": "hash.unraid.net", "internalport": "443", "internalprotocol": "https", "remoteaccess": "yes", "servercomment": "Media server", "servername": "Kris-Server", "externalhostname": "hash.unraid.net", "externalport": "[redacted]", "externalprotocol": "https" } Response (HTTP 200): [] success

Yup, I'm an idiot. Thanks for the reminder. I certainly need it sometimes.

Since the most recent update my server can no longer connect to the network. eth0 and eth1 don't even show up as devices in ifconfig. I am not very savvy with the command line and could really use some help. I've attached the diagnostics file. kris-server-diagnostics-20220604-2012.zip

18x 4GB sticks for a total of 72GB (I just have a problem with empty slots)

I was on the verge of abandoning Unraid altogether before I found your post. You definitely saved me days of work. My server has been humming along nicely ever since.

I just found this post/guide from last year. My transfer has just gone from 4 MiB/s to over 100. Now hopefully I can transfer all of my files and then have a 12 hour parity sync at the end. Someone needs to make this a sticky!!

Yes, Disk 6 is brand new. Less than 200 hours on it now and almost no writes.

Seagate lists that model as perpendicular. Not that means they are any good. I was trying out the WD drives after my third failed Seagate.

Then wouldn't my performance improve when writing to the Seagate drives? It seems to actually be faster when it was on the WD. Now that it hit 1TB it has moved onto a Seagate disk and seems even slower, albeit not much slower. Is this just a normal function of calculating parity? Also, I'm assuming that read speeds won't be this problematic. What is the best way to test this before dropping $90 on Unraid? I've already spent a week trying to set this up, I'd hate to waste the money too only to realize I'll be better off going back to Lubuntu with the old RAID card.

I understand I won't be getting the best performance from the WD drives. I'm even okay with half the speed. Before switching to unraid I had these exact drives in a traditional RAID 5 connected to the onboard P410i and was getting 40-60 MiB/s. Not to mention at present it is writing to a Seagate drive which is perpendicular recording. I'm not expecting blazing fast speeds. Just better than 4 MiB/s. Once I'm finished dumping all my existing media on the array I'll be adding a TiB of SSD cache and will hopefully never be moving this much data to the array again so SMR should be a non-issue. I just didn't want to hit the SSDs when I have 8TiB of media to get everything started. At this rate my trial license will run out before I get a chance to actually test my PLEX setup with all my data.

So I started up a large transfer again. The speed starts out great at about 130 MiB/s. After about 2 minutes it's down to 30 MiB/s. At the 5 minute mark it's down to 20 and just keeps sliding downward. By 10 minutes it's at 4-5 MiB/s and just stays there.

Everything seems normal with one obvious outlier that is oddly fast. Not sure what this means.

Maybe slightly faster? Seems pretty much the same to me. But copying to my desktop I get about 130 MiB/s from the USB drive.

I saw someone mention on another forum that HP RAID cards disabled the internal cache of HDDs connected to them. Is this possibly correct? I know the raid card had its own cache. Would this cause horribly slow writes? If so, how would one go about re- enabling the internal cache on the drives?

No, disk one had a lot of reallocation errors the first time I attempted to copy files over. I stopped the transfer and replaced the drive thinking it might be causing some issues. It was several years old. Rebuilding the drive only took 12 hours and that was completed before trying to transfer more files to the array.

Apologies, I'm very new to this. Diagnostics attached. kris-server-diagnostics-20210915-0633.zip

This is my first time using Unraid. I am currently running version 6.9.2 on an HP ML350 G6. I have 2x Xeon X5675 CPUs and 72 GB ECC RAM. I have setup an array with 8 2TB SATA disks 2 of which are parity drives connected with an LSI 9207-8i in IT mode. I installed the plugin for Unassigned Devices and am using the Krusader Docker for initial file transfer. I have the PLEX Docker installed also but it is not currently running. Now that I have the specs out of the way, I am attempting to transfer my media backup from an 8TB USB 3 HDD to my Unraid Array. I am now getting speeds of 3-5 MiB/s with an expected completion of 6 days for a 2.9 TiB transfer. Am I doing something wrong? I see other posts complaining about 40 MiB/s. I would be thrilled to hit that mark. Any assistance would be appreciated. Thanks.