Frank1940

The Marvell chip sets have been an ongoing problem with Unraid for quite some time. Their use is not recommended as they sometimes work and sometimes don't work. Yours (apparently) works if you disable VT-d. If you require VT-d to be enabled, you will have to replace it. LSI based SATA carders are recommended for use with Unraid these days.

I would be looking at the cat5E/6 cable and then at the port on the switch. You might also read the manuals for the MB/Lan Card and the switch to see if the lights on the RJ45 plugs indicate what speed was negotiated. If you find nothing there, post up a Diagnostics file ( Tools >>> Diagnostics ) in a new post.

You may be suffering from one of the problems described in this thread: https://forums.unraid.net/topic/25064-user-share-problem/?tab=comments#comment-228392 Start with this post and continue reading as the next few posts describe other gotchas!

Parity 2 and Parity 1 combine to provide protection for failure of up to two disks without data loss. The use of either Parity 1 or Parity 2 provides for the loss of one disk without data loss.

This would be about right from my observation for USB2. USB3 should be considerably higher. Also File size has a lot to do with copy speeds.

Most (if not all) 4K TV's will do their own upscaling. While I don't run Plex, I would suspect that its upscaling algorithm is not superior to that in the TV. (If it does a cruddy job, you should not have purchased that TV in the first place!) Transcoding is usually required when you are feeding high resolution stream to a low resolution device. A old rule of thumb is 2000 passmarks per stream. The only other time you might need transcoding is if you had some video with a non-standard resolution that the TV could not handle. Your choice of CPU is fine. The comment was made more in reference to the lack of a SSD for cache then about the CPU. (Having a cache drive for VM's and Dockers makes for much better performance!) That CPU has a lot of computing power and you should be quite happy with it for a long time. For a simple server, it would be overkill but IF you want to have the option of running VM's, you will be quite happy with this choice. Since you have it--try it and see if it works,. I pointed it out in case you had not purchased Memory yet. Since that MB supports ECC, a lot of Unraid Gurus recommend using it when possible. (Most non-server MB's do not support ECC and most server MB's do support it. A bit of logical thinking brings one to the conclusion that the folks who run the server farms want ECC in their servers!) If you run it as a Docker, technically not. You could setup a separate VM and run Plex on that VM but I don't know of a reason why you would want to do that. One more thing, if you are thinking of setting a VM, you will probably need a GPU at that time as I don't believe that that MB/CPU combination supports on-board video.

From your screen capture of the Main tab, nothing is being written to any device on the array. (I am assuming that it was taken when this activity was going on.) Now, 7.2Mbps is not very fast (for a data transfer). I would suspect that it is GUI activity. Did you have more than one GUI screen open? Were you 'watching' the preclear progress on those three drives?

There are two three points I wish to make. First, you don't have an SDD on there for use as a cache drive. (Having E5-2680-v3 leads me to believe you are going to be running a VM at some point.) Cache will also speed up transfers to the array. IF you are concerned about lack of data protection, get two SSD's and set a protected cache pool. Second, Unless you actually have 50TB of data ready to load on the system, reduce the number of 6TB drives back to a more reasonable level. It is so easy to add disks to Unraid. You shut the server down, physically install the disk, start the server up, assign the disk to a data slot and start the array. Another reason to spread out disk purchases to prevent buying into a lot of disks with poor manufacturing quality level. Spreading the purchases out also reduce the risk of long term shipping damage due to exceptionally rough handling of the delivery package. (Plus, disk prices for a given size tend to come down over time. Waiting a year of two could result in a significant saving.) EDIT: You should be ordering ECC Memory as that MB supports it.

You can copy from a User Share to a User Share. You can copy from Disk Share (i.e., /mnt/diskX) to Disk Share. What you never, never, never want to do is to copy from a User Share to a Disk Share (or vice versa). In your case, it is safe to Copy(/Move) from Disk1 to Disk2. Actually the 'Move' operation is probably your best choice in your situation as having duplicate files in a share can result in a some puzzling issues!

You may be right in a few cases but generally, the volume of air moved pushing against a head is virtually equal to the volume of air moved when sucking against a 'vacuum'. Air exiting under pressure will still seek the path of least resistance. The volume of air moved by most fans is often a couple of multiples larger at high speed than at low speed against the static pressure but so is the noise. The other disadvantage against blowing the air in at the rear is that it is first heated by the electronics back there. The electronics can tolerate higher temperatures far better than hard drives. Furthermore, look at the commercial servers, they all suck the air in over the drives and exhaust it out the back. ...And they basically use the same type of drive enclosures as these NAS do.

You need to address your cooling issue now. With servers, you want the air to enter the case (most likely from the front) and flow over the hard drives and exit through the rear of the case. You also want fans that can move air against static pressure. (These generally make more noise but there are some that can both move air and are quiet(er). But they cost more...) Don't leave the side of the case open thinking it will keep the drives cooler. It won't unless you have twenty inch box fan blowing right at them. Air flow is your only friend in cooling hard drives.

I would replace that parity drive immediately. With over 13,000 remapped sectors and over 2400 sectors needing to be remapped, it is a disaster waiting to happen Even if you could get it to remapped those 2400 pending (and, apparently, completely unreadable) sectors to remap, when would the next one popup? Probably right in the middle of a rebuild of another disk which would cause that rebuild of that disk to fail.

CRC errors are not Drive errors. They indicate a failure of the CRC check of the data being transmitted over the SATA data cable between the hard drive and SATA controller where the CRC is verified to determine that there was no corruption in the transfer. (By the way, this number is permanently stored on the disk and can not be reset. Monitor to make sure it does not increase.) It is usually caused by a bad SATA cable or a loose connection of the cable. The data on the disk is most likely fine. The data will simply be retransmitted until it is received without failing. You should post up a Diagnostics report Tools >>> Diagnostics so that someone can have a better look at the parity drive. (Personally, I would replace the parity disk, rebuild parity on that new disk. Then I would have a better look at the old parity drive to see what its issues are. With the limited data that I see here, I would suspect that it is toast.) EDIT: By the way, 227 days between parity checks is much too long. You should be doing it (at least) monthly. If you had not done it, you should also setup the notification system and have it sending you status reports on the health of your server.

Both looked good to me.

I am not sure why you got this result but I want to point you to this WIKI: https://wiki.unraid.net/Check_Disk_Filesystems#Checking_and_fixing_drives_in_the_webGui It is possible that you were using a -n option which means that the disk was only analyzed and not fixed. Another point. Be sure to capture all of the output from any command (along with the actual command you ran to get those results) and post it with any queries that you have. In doing file system repairs, it is always better to ask a question than to really screw something up because you are not sure of what to do next. I am one who does not have much experience in file system repairs (Knock on Wood!!!) but if you can't figure something out I will ping a gentlemen who has help many folks through this process. EDIT: While posting up command lines and the result of command lines, please use the code format (the </> icon on the formatting bar of the reply box). It is so much easier for us to recognize what-is-what if it is formatted.