bmfrosty

If it's as simple as that, it's as simple as that.

I recently replaced most of my drives and am in the process of migrating data around to reformat and I've noticed my speeds rebuilding were slow, and copying data currently is about 73MB/s with unbalance. I'm using on-board SATA ports, and I was wondering if that may be why I'm getting such bad speeds. What are the usual culprits?

Just hit this last night and it kept me up late thinking about it. 5 new drives for a ZFS array. I'm going to have to rethink my plan. I had no expectation that this would be a restriction. I guess every share that's not getting incoming writes goes on the ZFS for now.

Did a preclear to the new drives for burn-in over the past 6 days. Added a new pool with 5 devices set to raidz2, compression off, autotrim on, enable user share assignment yes, and started my array. I then went to to the main tab, array, clicked the checkbox by format with the 5 drives showing, clicked the checkbox, acknowledged the warning, and hit format. I get started formatting, then it goes back to started. Pool shows a bad device. Am I doing something wrong here? edit: ooh `Nov 14 20:54:57 maxi root: wipefs: /dev/sdb: failed to erase dos magic string at offset 0x000001fe: Operation not permitted` ``` root@maxi:/dev# for i in sdb sdc sdd sde sdf ; do dd if=/dev/zero of=/dev/${i}1 bs=4096 count=4 ; done dd: error writing '/dev/sdb1': Operation not permitted 1+0 records in 0+0 records out 0 bytes copied, 9.1293e-05 s, 0.0 kB/s dd: error writing '/dev/sdc1': Operation not permitted 1+0 records in 0+0 records out 0 bytes copied, 3.7661e-05 s, 0.0 kB/s dd: error writing '/dev/sdd1': Operation not permitted 1+0 records in 0+0 records out 0 bytes copied, 3.8583e-05 s, 0.0 kB/s dd: error writing '/dev/sde1': Operation not permitted 1+0 records in 0+0 records out 0 bytes copied, 9.7745e-05 s, 0.0 kB/s dd: error writing '/dev/sdf1': Operation not permitted 1+0 records in 0+0 records out 0 bytes copied, 3.5366e-05 s, 0.0 kB/s root@maxi:/dev# ``` same for the raw devices I need to come back to this in the morning. Solved. Had to clear it with the preclear script.

I've been (fairly passively) using unraid for about a decade, and my oldest drives (3/5 in my array) have hit the 5 year mark, and I'm thinking that it's time to replace them. In the process, I'm considering moving to a ZFS pool. My current setup includes a 2-drive SSD Cache using BTRFS and a 5 drive array using XFS on the drives with 1 parity drive. My goals are to increase storage and improve failure tolerance and reduce the occurrences of having to use unbalance to shift around files to make space. My current thoughts are to build a 5-drive RAIDZ2 ZFS pool to take place of my array (mostly) which should bring my usable space in the pool to ~56TB vs the usable space of my array of ~44TB. I'll keep the two newest disks in the array as still an array, and will keep (usually newer) frequently accessed files there. ZFS pool will keep lighter touch files. I guess my questions are - Does what I'm thinking about here make sense? When not being actively access for a time will the drives in the ZFS pool spin down? What might I not have thought about yet that I may need to think about?

Just posting because I had been having an issue where where privoxy was very slow. Didn't matter, I only used it to log into a site or two, and qbittorrent worked well. That changed about 3 or 4 days ago, and I couldn't get entries the RSS feed working on this specific site that I log into with qbittorrent. Turns out my list of dns resolvers had too many bad entries. I switch to the cloudflare resolvers and suddenly everything is fast and the RSS works. Go figure.

I switched from a Vega 56 to a 2060 today and I appear to be having better luck. I didn't think that I'd have that on a linux system with nvidia. I need some DX10/DX11 games to test that are known to work. Anyone want to throw out a couple of suggestions?

A bunch of things are now working. The container definitely works, I can VNC in and it definitely works, and I can play games that are DX9 and they work. I was able to launch Saints Row IV and Portal 2. What's not working is anything past DX9. I'm running a Vega 56, which does have DX12 support, so I'm expecting it to be a drivers problem. I've tried GTAV and Darksiders (Warmastered Edition), and GTAV complains that my hardware doesn't support DX10/11, and Darksiders complains that the hardware doesn't support DX10. I'm not sure how to troubleshoot this. Older games *tend* to work as long as I'm running them through Proton and not relying on the Linux version. Can anyone point me in the right direction?

Just got it working on my current motherboard (From 2014!). Turns out that my motherboard doesn't support PCIe bifurcation, so I had to use two cards in two different PCIe slots. I like the idea that btrfs will let me freely switch between jbod and raid1 as long as I have free space. I'm doing that to download a couple of very large things, and then I plan to turn it back to raid1 assuming that I can manage what I think it does. Being very careful about taking backups before I do operations like that though. I think the bridge is going to allow me to do some fancy things. 2x1tb is great for now, but I'd like to have raid1 and 2tb of usable space some time in the future with the option to double it using raid5 and still keeping the protection. For right now, this is a big improvement over the 600gb SSD that I've had in there since about when I built the thing in 2013 originally. 2014 motherboard didn't get installed until 2020 IIRC. Basically I need to understand what lanes are in use and which ones aren't. I think the 5900X has 24 lanes, 4 of which are for the bridge that then in turn provides 16 more lanes (that all have to feed through the 4, but that's alright for this application, I think), but I need to look at some docs and figure out how many m.2 I can effectively have while still getting enough SATA ports on the board. Fun, huh? A quick read shows that I can get one m.2 m-key slot on the motherboard direclty off of the CPU and one off the x570. Also there's an x16 (that's really x4) PCIe slot off of the x570 that I can get a 3rd m.2 slot out of. There's also an x1 slot that I can get a couple of SATA ports off of as well, I think. If I'm keeping my GPU (I am for steam-headless) then that's about as far as I can go on this motherboard and processor, which is just fine by me.

I'm heading very much in this direction. I'm just questioning whether I can get two cache m.2 drives in. Still doing the research, but thinking that I may buy the 5900X while it's still on sale even if I end up with a different motherboard.

Got everything working yesterday. Had to change a kernel parameter and had some conflict that I had to solve between this container and a VM - (I didn't fix it, I reinstalled the chart, which caused the problem to go away), and I updated this AM, but now in noVNC I get a popup that I didn't have yesterday: noVNC encountered an error: The play() request was interrupted because the media was removed from the document. https://goo.gl/LdLk22 I saw this occasionally before when I was testing and got the VNC for my container, so I don't know if it's a regression here or elsewhere.

No luck on my side, but I'm on 6.9.2. I'll upgrade to 6.10 tomorrow. No luck there either, but found a problem of some sort in dmesg- [26167.409463] pcieport 0000:00:03.0: AER: Uncorrected (Non-Fatal) error received: 0000:02:00.0 [26167.409473] pcieport 0000:02:00.0: PCIe Bus Error: severity=Uncorrected (Non-Fatal), type=Transaction Layer, (Requester ID) [26167.409476] pcieport 0000:02:00.0: device [1022:1471] error status/mask=00100000/00000000 [26167.409478] pcieport 0000:02:00.0: [20] UnsupReq (First) [26167.409480] pcieport 0000:02:00.0: AER: TLP Header: 34000000 03000010 00000000 84288428 [26167.409489] [drm] PCI error: detected callback, state(1)!! [26167.409503] pci 0000:03:00.1: AER: can't recover (no error_detected callback) [26167.409511] pcieport 0000:02:00.0: AER: device recovery failed Repeats ~1000 times per second.

I found that 4.3.x versions were memory hogs in docker. The move to 4.4.x reduced the memory usage down significantly.

On the RSS Feed issue. It's a recent change to QT6. Info on it: https://github.com/qbittorrent/qBittorrent/issues/16879 EDIT: Easy fix is to change your repo to: binhex/arch-qbittorrentvpn:4.4.2-1-01 Assuming that it's fixed in a few weeks, you can change it back to binhex/arch-qbittorrentvpn

Great. Makes things easier for me. Very good to know. I should be done during the weekend then.

If I'm mistaken, then I'm mistaken. My understanding (definitely old) was that a preclear would reduce the amount of time spent during the rebuild of the data from the 3TB drive to (and I'm forgetting here) clear out all the data past the 3TB mark or (and now that I'm writing it out it seems unlikely) prevent the old parity data on the old parity drive being used to update the parity information on the new parity drive since it is readable data in TB 4-10. Am I correct in needing to clear the first few sectors of the old parity disk so it's not recognized by Unraid as part of the array?

I have an array with: 1 SSD as cache pool 1 10TB disk as parity 2 10TB disks as data 1 3TB disk as data. I am out of drive bays. I've removed the parity disk and replaced it with a precleared 14TB disk. I'd like to preclear the old 10TB parity disk and use it to replace the old 3TB disk. I think my steps are: 1) Plug the old parity disk to a USB enclosure 2) Clear the first few sectors - `dd if=/dev/zero of=/dev/hdx bs=1024 count=10` - to clear the preclear signature 3) Run the preclear script on it. 4) Shut down the array and power down 4a) Swap the 3TB disk for the 10TB disk 5) Power Up and allow the parity to rebuild the 10tb disk Is this the best method? Am I missing a step somewhere? Am I completely off base?

I can't connect to other hosts in my lan from the binhex-qbittorrentvpn container. I've been using this for a while (I am currently using PIA/Wireguard for my VPN) and have recently set up jackett as well using the binhex-jackett image. Jackett itself seems to work just fine. I think tried to add a Jackett RSS feed to qbittorrent. It can't connect, so I set up another Jackett host on a raspberry pi on my network. I can't connect to it either from the container. I can ping it from the container. I can not curl it from the container using hostname, FQDN, or ip address. I can ping it from the unraid host and I can also curl it. I'm usually self sufficient at these sorts of things, but at this point I feel it's better to ask the thread as I assume this is an already solved problem (or unsolved for various reasons!). Is there a short answer? Am I best off just setting up jackett on linode or something - not that I'm terribly interested in doing so? Thanks!

It's coming time to upgrade my motherboard again - to add more RAM than the current one supports, and with that will come a new CPU. In the process I've been thinking about how much easier it would be to buy a small desktop off the shelf, add a SFF-8088 card, and just plug it into a DAS instead of opening up my current case and doing all the work in there. I've been doing some googling on jbod chassis that support SFF-8088 (or optionally SFF-8644) and they're not too hard to find - from multiple vendors - including synology. I was wondering if anyone else has gone through this process and could speak to what sort of problems and caveats they've run into doing this sort of transition.

Trying to hit an rss feed that's also being served via unraid and having no luck. Is there anything to do to let rtorrent hit my internal network?

Is there a template for this? I'd like to give it a try, but don't want to make too much of a project of it just getting off the ground.

BLUF: Have UNRAID running with this hardware, looking to run a Windows VM for steam remote play. I've spent a little while playing in the GUI and don't even see GPU passthrough. Don't really know where to start. Looking for someone to point me in the right direction. This is my old gaming rig from 2014, that I started running POP-OS on, and replaced the video card on at one point. I've switched to a System76 Laptop for my daily driver, and moved motherboard/cpu/ram/gpu into my unraid case for a nice upgrade from my old CPU. I was hoping to run a Windows VM and pass through the GPU so I could do Steam Remote Play. I've been poking through the VM setup GUI, and I don't even see the GPU as available for passthrough. I would love it someone could tell me where to even start. Can someone point out where I should start on this?

So far, I'm either looking right past it, or it isn't there. Do you have a specific link? That took me an embarrassingly long time to find.

Is there any way to stick to stable versions of the client? I'm connecting to a tracker that appears to be rejecting me because I'm using 2.0.4.dev23 on this. I'm inquiring as to whether I'm correct on this from the tracker's forums. I'm currently transitioning from my laptop which is on 2.0.3 and is working just fine with said tracker.

I stopped and started unBalance, and it was the newer version. It worked on my previously problematic sample. Thanks for the fix!