stevenxl

Thanks. At this point I think this is just the performance of these drives. I cheaped out on the whole setup thinking it would be enough. I think they are frankly, but the drives are crap. Not to high-jack my own thread, but can you recommend one or two product lines for HDDs in the 8TB range that will give me better speeds than what I am seeing.

I thought that SMR only mattered in a RAID configuration. I don't have RAID enabled, thanks to the way that unRaid OS does the parity stuff. Here are the results on running hdparm on disk2 and 3: ./disk2: Timing cached reads: 39360 MB in 2.00 seconds = 19712.98 MB/sec Timing buffered disk reads: 466 MB in 3.00 seconds = 155.08 MB/sec ./disk3: Timing cached reads: 39396 MB in 2.00 seconds = 19729.99 MB/sec Timing buffered disk reads: 594 MB in 3.00 seconds = 197.75 MB/sec

For what it is worth, I ssh'd into unraid and ran `dd if=/dev/zero of=/mnt/disk2/archive/testfile bs=1G count=5` and got 28MB/s. This rules out a problem with SMB, WiFi, etc. I ran `lsusb -v` and confirmed that that the device was recognized as USB 3.1: Bus 004 Device 002: ID 1c04:0013 QNAP System Inc. TR-004 Device Descriptor: bLength 18 bDescriptorType 1 bcdUSB 3.10 `hdparm -tT /dev/disk/by-label/disk1` gave me: ./disk1: Timing cached reads: 39324 MB in 2.00 seconds = 19694.33 MB/sec Timing buffered disk reads: 620 MB in 3.00 seconds = 206.50 MB/sec And `smartctl -a ./disk1` is a bit useless as the drives don't have any SMART capabilities: root@unraid:/dev/disk/by-label# smartctl -a ./disk1 smartctl 7.4 2023-08-01 r5530 [x86_64-linux-6.1.79-Unraid] (local build) Copyright (C) 2002-23, Bruce Allen, Christian Franke, www.smartmontools.org === START OF INFORMATION SECTION === Vendor: QNAP Product: TR-004 DISK01 Revision: 6111 Compliance: SPC-4 User Capacity: 8,001,563,222,016 bytes [8.00 TB] Logical block size: 512 bytes Physical block size: 4096 bytes LU is fully provisioned Logical Unit id: 0x3051323433423030 Serial number: 5132343342303035 Device type: disk Local Time is: Wed Sep 4 23:51:35 2024 CDT SMART support is: Unavailable - device lacks SMART capability. === START OF READ SMART DATA SECTION === Current Drive Temperature: 0 C Drive Trip Temperature: 0 C Error Counter logging not supported No Self-tests have been logged

I am transferring 1 2GB file, as this mirrors my usage of the system much more closely than many small files.

I am wondering when to capture the diagnostic file. Should I do the tests you've outlined and, after each test, download the diagnostic file?

Hi folks, My current setup is that I have a 4-bay NAS (QNAP TR-004) connected to my ASUS - Zenbook 14X 14.5" laptop over USB 3.2 Gen 1. The drives in the DAS are Seagate BarraCuda 8TB Internal Hard Drive HDD – 3.5 Inch (ST8000DMZ04/004). There is a bunch of stuff wrong with this setup. The drives are not great drives (something about SMR vs CMR), connecting via USB is a bad idea, the laptop is not a real server, etc. etc. However, I still do not understand why the speeds I am getting are below 20 MB/s. The drives support 6 Gbps but the max sustained transfer rate is 190 MB/s, USB 3.2 Gen 1 supports supports 5 Gbps, my laptop supports USB 3.2 Gen 2 with speeds of 10 Gbps. The limiting factor here should be the drives at 190 MB/s; cut that in half and I'd still expect 100 MB/s. What am I missing? I guess another link in the chain is my WiFi. This is all happening over LAN, but the WiFi is still involved here right? That's able to do at least 60 MB/s. The computer is this (https://www.asus.com/us/laptops/for-home/zenbook/zenbook-14x-oled-q420/techspec/). The DAS is this (https://www.qnap.com/en-us/product/tr-004) The drives are here https://www.seagate.com/www-content/datasheets/pdfs/3-5-barracudaDS1900-11-1806US-en_US.pdf, though I can't find my exact model.

Oh - it looks like this post suggests that the built-in nginx can't be used for anything else:

This doesn't work for me. I don't think that the files in `/etc/nginx/conf.d/` are automatically being picked up. Am I missing something?

Hello. I have a subdomain that is of the shape `unraid.<mydomain>.cc`. The DNS record is set to point to my unraid server - i.e., 192.168.50.231. This works fine because the "native" nginx on unraidOS listens to port 80 / 443 and routes requests to the unraid management app. However, I want to create `webtop.<mydomain>.cc`, and have that routed to 192.168.50.231:3000. How would I go achieving this? I don't want my DNS record to have my WAN IP as is necessary for publicly-facing applications.

It turns out that the `.pem` file needed by unraid has to be a "bundle" - i.e., it includes both the cert and the key. This is easy to do by concatenating both files with `cat`. Now, you do have to manage updating the cert whenever `swag` updates it, but this is easy enough to do by having a script run on a regular basis.

Hi folks, I am using the `swag` container to get HTTPS working for publicly facing services. This so far works fine. I have: 1. `ddns-updater` able to update my WAN IP on Cloudflare 1. I have `swag` able to generate a wildcard certificate I have exaclidraw working on `draw.<mydomain>.cc` with HTTPS. I am trying to do the same with internal-only services. I tried to use the certificate generate by `swag`, which is available in `/mnt/user/appdata/swag/etc/letsencrypt/live/<mydomain>.cc/fullchain.pem`. In order to avoid having to manage this certificate manually, i was going to symlink it to `/boot/config/ssl/certs/unraid_unraid_bundle.pem`, but that leads to an error. What else can I try? Just copy the letencrypt certificate locally and copy it again when letencrypt updates the certs?

That is fair. Thank you!

So an exclusive-share will by-pass the FUSE filesystem. Will writing to /mnt/disk1 do the same?

Hm... the parity check was ran less than 30 days ago. You think the USB speed would have degraded in that time? I think I'll try a cache -> disk transfer to see what happens when FUSE is by-passed.

I understand the concern regarding USB connection, but I was doing the parity check on that same setup and it was doing 60MB/s. What would explain a drop to 25MB/s?

Thank you for this piece of advice. Things did speed up significantly: > 13,342,405,607 100% 21.66MB/s 0:09:47 (xfr#2, to-chk=0/3) I'm attaching the logs here (taken during transfer) in case there is a way to get better performance. Something still seems off, but maybe this is my expectations being wrong? P.S. Given how much fun I am having with unRAID and how great you have all been at support, I am basically at the point of buying a license. unraid-diagnostics-20240614-0939.zip

Hey folks, I am transferring a file from my cache (https://www.micron.com/products/storage/ssd/client-ssd/2400-ssd - the 512GB version) to my array. The data transfer rate is seems very slow: > 9,394,687,673 100% 13.77MB/s 0:10:50 (xfr#1, to-chk=0/2) 13.77MB/s seems slow. By contrast, the parity check was showing an average of 60MB/s: > Duration: 1 day, 13 hours, 19 minutes, 49 seconds. Average speed: 59.5 MB/s What could be wrong with my setup? I am transferring a large (10GB, roughly) file using `rsync` from my NVME SSD to my HDD arrays. Any thoughts?

It turns out that I can load an HTTP file: https://support.plex.tv/articles/using-an-xmltv-guide/ I just uploaded the guide to somewhere publicly accessible and pointed Plex to that. Worked extremely well!

Hello everyone. I am trying to set up Live TV on Plex. The source of the TV Streams is an IPTV, so I downloaded and set up xTeVe. I am able to watch the channels fine, but when adding a "Tuner" that is really the xTeVe server, I was unable to load the EPG / XMLTV file into Plex. (I was able to load it into xTeVe, so I wonder why I have to load it into both apps, but I digress). Where should the file live in Plex so that it can be found?

I am trying to follow Space Invader's video on how to set up my shares (link below). The advice (for a media server setup) is to set up programs that download torrents or other files to download to a cache-enabled share. Then, move these files from a cache-enabled share to a share that is not cache-enabled. However, if I move the files without letting the torrent client now, then the torrent client will stop seeding the files (since it won't have them). How can I deal with this issue? In case it matters, I am using `qbitorrent` as my torrenting client (though I prefer rtorrent, there's not rutorrent in the app store).

Did you ever find a solution to this? I too am looking for this.

So I was connecting the DAS to a USB-C hub and then connecting the USB-C hub to the USB-C port on my computer. I did this because I thought USB-C was faster than USB-A. What I failed to understand is that the USB-A and USB-C are references to the port type. I need to look at USB 3, etc. which is what determines the data speeds. All of that to say that I fixed this by plugging in the DAS directly to the USB 3.2 Gen 2 port on my computer. Now I am getting much faster speeds. I stopped the parity-sync, to do some other work, but it was reaching 40+ MB/s and the time was going down to at most 1 day. Thank you @JorgeB.

@kilrah point taken on the laptop, but let's skip that for now. I'm looking for advice on the "thing" that holds my hard drives. The product I linked above does use SATA-connected drives, but that DAS is connected to my server via USB 3.2 Gen 1. Wikipedia says: USB 3.2 Gen 1×2 – new, 10 Gbit/s signaling rate over 2 lanes using 8b/10b encoding (nominal data rate: 1000 MB/s). Is the DAS enclosure the issue? Because it is SATA and the nominal data rate is 1000 MB/s, doesn't seem that it should be that slow.

Hi @jorgeB. Thank you for taking the time to respond. My drives are on a DAS - this is the model: https://www.qnap.com/en-us/product/tr-004. Is that the limiting factor here? What alternatives would you recommend? What should I have been looking for? Sorry - real newbie here.

Turns out that the permissions on the files / directories are wrong. I changed the files to 666 and the directories to 777: `find . -type f -exec chmod 66 {} \;` `find . -type d -exec chmod 777 {} \;`