Yekul

I have made a stupid mistake which FORTUNATELY I managed to snag because of dumb luck (my coral TPU inferences were very high). Could you please take a look at below and provide some suggestions. I think i've tidied things up a bit to prevent further issues but would appreciate any extra insight without too much 'the world is ending' where possible. I closed out some dockers and CPU was being pegged to 100%. Checked logs and a flood of remote connections to port 22 (SSH) which I -stupidly- left open after some issues when moving house and forgot about. Now I have disable SSH remote access, and it does seem as though all the incoming connections were on port 22. Had to end a task to get the CPU to return to normal. My password is very complex fortunately, so the odds of it being beaten by a login attempt like below seems... low? Or is there actually many more which the logs just don't record? I do have my router forwarding connections from the wan to my unraid ip address for reverse proxy domains. I assume this is OK now SSH is disabled? Of course the reverse proxy has a unique username/password for external access. I will say, i'm surprised there isn't a default action to block logins after x amount of failed attempts with a set time. However I do note the IP address is constantly changing. My regular login for Unraid is done through Wireguard or Unraid Remote. Example of incoming connections below: Jan 13 15:44:30 sshd[28347]: Connection closed by authenticating user root 170.64.214.0 port 46188 [preauth] Jan 13 15:44:30 sshd[28431]: Connection from 170.64.214.0 port 46194 on 192.168.1.99 port 22 rdomain "" Jan 13 15:44:32 sshd[28444]: Connection from 24.144.80.196 port 56214 on 192.168.1.99 port 22 rdomain "" Jan 13 15:44:32 sshd[28447]: Connection from 101.251.197.238 port 35040 on 192.168.1.99 port 22 rdomain "" Jan 13 15:44:32 sshd[28444]: Invalid user wangxh from 24.144.80.196 port 56214 Jan 13 15:44:32 sshd[28444]: pam_unix(sshd:auth): check pass; user unknown Jan 13 15:44:32 sshd[28444]: pam_unix(sshd:auth): authentication failure; logname= uid=0 euid=0 tty=ssh ruser= rhost=24.144.80.196 Jan 13 15:44:33 sshd[28474]: Connection from 121.46.20.110 port 59892 on 192.168.1.99 port 22 rdomain "" Jan 13 15:44:33 sshd[28431]: pam_unix(sshd:auth): authentication failure; logname= uid=0 euid=0 tty=ssh ruser= rhost=170.64.214.0 user=root Jan 13 15:44:35 sshd[28447]: Invalid user fang from 101.251.197.238 port 35040 Jan 13 15:44:35 sshd[28447]: pam_unix(sshd:auth): check pass; user unknown Jan 13 15:44:35 sshd[28447]: pam_unix(sshd:auth): authentication failure; logname= uid=0 euid=0 tty=ssh ruser= rhost=101.251.197.238 Jan 13 15:44:35 sshd[28444]: Failed password for invalid user wangxh from 24.144.80.196 port 56214 ssh2 Jan 13 15:44:36 sshd[28444]: Received disconnect from 24.144.80.196 port 56214:11: Bye Bye [preauth] Jan 13 15:44:36 sshd[28444]: Disconnected from invalid user wangxh 24.144.80.196 port 56214 [preauth] Jan 13 15:44:36 sshd[28431]: Failed password for root from 170.64.214.0 port 46194 ssh2 Jan 13 15:44:37 sshd[28447]: Failed password for invalid user fang from 101.251.197.238 port 35040 ssh2 Jan 13 15:44:37 sshd[28447]: Received disconnect from 101.251.197.238 port 35040:11: Bye Bye [preauth] Jan 13 15:44:37 sshd[28447]: Disconnected from invalid user fang 101.251.197.238 port 35040 [preauth] Jan 13 15:44:38 sshd[28431]: Connection closed by authenticating user root 170.64.214.0 port 46194 [preauth] Jan 13 15:44:38 sshd[28543]: Connection from 170.64.214.0 port 59152 on 192.168.1.99 port 22 rdomain "" Jan 13 15:44:39 sshd[28543]: error: kex_exchange_identification: read: Connection reset by peer

Just passing along info as I was plagued by this error message after changing my password today. Wouldn't let me back in on multiple browsers (firefox, chromium, edge). Cleared cache/cookies. Restarted PC. Tried 2FA, Tried alternative methods to 2FA. Nothing worked just kept locking me out for 10 more mins. Randomly somewhere in the middle Edge logged in, and when I logged out and back in to verify if it was indeed fixed then it stopped working all together again. Left computer alone for ~1.5 hours and now without changing anything I can suddenly log in again. Really have no idea what was going on. When it wasn't working, I would enter username/password as usual, sign in, have the 2FA popup come up, enter code, hit enter and then be bounced back to the forums page with address https://forums.unraid.net/login/?_fromLogin=1. However nothing would actually happen. It was as if it didn't even try to log in. Will provide feedback if it resolves or continues over coming days.

Probably a long shot, but any way to control the % of overall cores/threads tdarr uses? Was hoping to have a situation where you can specify say 90% usage of all cores rather than pinning and locking out an entire core/thread. Been having solid success so far with tdarr, so thanks OP.

Having a quick play with this, and the foundations look real solid and almost just what i'm after. I am curious though, any plans to add a manual scan button to the front page, and perhaps a 'start paused' type option? So we could select a single folder from the scan and start it before bed for example. I prefer to be more in control of what is happening is all, so being able to scan and encode a folder at a time without -everything- being encoded automatically would be great. Plus then I can run it max speed when I know nobody will be using it. Haven't tested this part yet as running a test encode first, but assuming the advanced options is where I could change the video quality settings under 'customer ffmpeg options' to whatever I prefer? And I could only add video preferences and audio would pull the info from the 'audio encoding' tab?

Nevermind figured it out! And indeed, everything is working perfectly again on the older image. I swapped back to 'version-1.22.0.4163-d8c4875dd' with linuxserver.io image and that did the trick for me. Unsure if it's just me, but might be worth mentioning to save some people headaches should it be an actual issue.

Hrm thought this would be easy, but can't find a place anywhere in the linuxserver.io package to enter the tag so guessing I need to manually add it as a new variable, anyone more knowledgeable than myself mind sharing how to set the specific tag to use please?

Yeah suspecting it may be an issue with 10000 series? I dunno, have to spend some time getting logs for it I guess. Could have sworn I was watching a video file with it transcoding perfectly fine just the other day though, wonder if I can downgrade easily from latest version to test that theory...

Hrm, anyone else having issues HW transcoding 4k HEVC HDR content? My 4K HEVC content is hardware transcoding perfectly on my intel 10400, but throw HDR into the mix and it just reverts back to software. I have the HDR tone mapping option enabled of course.

Ok great to know thanks. And should I always use transfer over copy? I'm used to Windows thinking where if it fails mid copy it sometimes vanishes into the abyss so always tended to do a copy and hten delete the old data. Though I suspect the duplication of data in the array may be the reason for the transfer vs copy now i'm typing it out?

Phew, started to freak myself out a bit. Thanks for the clarification. So I currently have a 4 drive array (3 storage + 1 parity) but migrating from a Windows based setup so have been doing a hard drive shuffle during transfers as it didn't have the parity setup (just used off site backups). Ideally once all the files are onto these two drives, I can then wipe and add the third data drive to the array as i'd like to have specific shares assigned to specific drives. The logic being if I set two drives to shared media I can limit spin ups on the other disk which is largely used just for backups, software installers, things I use a few times a week not multiple times a day etc. When I do those transfers, they will indeed be within the array. In those instances I should only transfer (not copy?) from disk 2->disk 3, correct? I use Krusader which is passing /media to /mnt. For within Krusader it would be /media/disk 2 transferring to /media/disk 3. Is it ok to make the corresponding folder if it doesn't yet exist in disk 3? Or should I change the shares to be disk 3 only and drag a file onto the share to get it to populate a folder itself first? Had issues with permissions in the past with various things and i'm totally new to Unraid so don't want to stuff anything up at the shares level is all.

So have been migrating data from unassigned devices disk drive to the disks in the array on a new build. However, just stumbled upon multiple threads where people have been very clear telling users not to copy between disks/shares etc as it ends in disaster. My question is, does it apply for unassigned devices? I originally copied my files from unassigned to disk 1, and as that was close to filling I swapped to copying (yes copy not move) my files from unassigned to a SHARE. With the logic being, I wasn't sure if the OS needed to be the thing to create the folder on disk 2 or if I could do it manually. No worries I thought, as it's using high-tide so if I just copy to the share instead it will automatically make a folder and start piling stuff onto disk 2 (and it did). Now i'm freaking out a bit after reading all the threads. Appreciate any input and words of advice? Or is it all just related to disk->disk copies and not unassigned devices copies?

Tentatively solved. Downgraded to confirm it wasn't a 6.9.1 issue. Still present. Then found C1 state enabled in BIOS, since disabling appears to be holding steady. Likely would have stopped being an issue when I swapped to a new intel setup in a few days time anyway, but posting a reply in case others stumble on this using older hardware like mine (Ryzen 1700, AB350M).

Addition to the above, I just tested the power button 'graceful' shutdown when the system is responsive still and it works completely fine. Shuts down correctly no problems at all.

Ok so I tried to do this, it doesn't seem to be persisting through restarts. I have it set to ->settings->syslog server: local syslog server: enabled local syslog folder: app data mirror syslog to flash: yes (temporarily to troubleshoot, not like it writes much anyway with my current setup) However the diagnostic file still clears. I hooked up a monitor and noticed when I hit the power button and it started the 'graceful shutdown' that happens it starts doing that fine then gets to 'Starting diagnostic collection...' and then just seems to hang. So I suspect that's why i'm not actually getting any useful log files? The syslog.txt file inside the diagnostics folder is only showing information since the restart, nothing before. Attached logs anyway fwiw, but really just shows recent restart and one of the failing drives (same thing happens without it attached btw, as I only just plugged it in). picklerick-diagnostics-20210321-1021.zip

How do I do this when the system becomes unresponsive? I thought the log files cleared on reboot. Which is the only way I can access it, by doing a forced reboot. I installed the Fix Common Problems plugin, but couldn't find the Troubleshooting mode previously mentioned in older threads (I guess this actively wrote the logs somewhere else constantly so they'd be stored on a physical drive and not RAM?).

So everything was going fine, started the extended offline test again. I could navigate to different pages, no unresponsive aspects at all. Then around ~30% or so through the test, same thing as before. Systems locks up and wont let me access anything. Can't SSH to server either. Happens with both drives tested, and one of them is brand new and passed testing no problems. Obviously not saying it's impossible for both to be faulty, but would be a hell of a coincidence. I just can't figure out what is going on, because unraid doesn't appear to power down or restart due to this.

Restarted (had to hard reset with power button) and the drive is available and the array loads up fine with the one disc. Not really sure what's going on tbh...

Hrm yeah wasn't sure if the format was meant to be doing a preclear type situation where it zeros the drive or not. Any ideas where to look for issues? As I said, brand new to Unraid and just trialing to make sure it's fit for purpose before jumping over. Haven't updated BIOS in a very long time but it's an old board (ab350m) so doubt there'd be anything too pressing, not like it's an x570 etc that is probably still receiving a lot of software related updates that may effect one thing or the other.

Yeah I thought surely this can't be right. Thing is, it has happened with two different drives. Bit reluctant to power down right now as it's mid formatting, and I can hear it actually access the drive every so often when it's quiet. What's the best course of action here? So to clarify, using two different cables, two different drives, two different sata slots, I could try and swap motherboards, but this board was working fine previously and it has no problems actually accessing the drives. I can run the short SMART tests no problem for example, and track the progress through 10->100% status. The long test I walked away after it started, but it wouldn't connect to unraid via local GUI when I powered on my main PC this morning so I couldn't say if it had increased in % or was still locked to 10%.

So looking to jump into Unraid as i've found myself with a good opportunity to 'start fresh'. Downloaded it, opted in for trial to make sure it fits my purpose. One of my drives (the only SMR drive) reported a slew of SMART errors so I thought i'd do a bit intense scan to see what's going on. Went to bed, woke up to check it via web gui and it can't connect anymore. Ok weird I thought, it must have crashed. Was completely unresponsive, would just do the standard time out you get with no connectivity. So force restarted. Everything comes up fine of course, there was no pools or anything yet anyway. But in the disk profile page is lists the SMART scan as being interrupted. So check some other drives, add my first to the pool and begin the function of adding to pool and then starting the format. Again, it becomes -completely- unresponsive. No other pages will load at all. So I guess i'm just curious, is this normal? It doesn't give a progress bar or anything, all I have is 'started, formatting' and then I have to keep coming in and hitting refresh to see if it finishes? And if this is normal and it locks you out of the UI intentionally, follow up question; does this mean the shares themselves go offline every time something has to be formatted to be added to the pool etc? 8TB is the smallest drives I have so that's a lot of downtime if so particularly during the setup process where i'm likely to add/remove a few drives over the next few weeks.

So assuming I use a dedicated GPU (something like a GTX1060 etc) and I wanted to passthrough to a VM for some light gaming, can I also use it for a Plex docker transcoding card? Or once it gets passed through to a VM is that it? Just wondering if it's possible to have it be used by a wide array of apps or if it's a 'one and done' situation? Probs a stupid question, but just trying to ensure I have hardware compatibility for what I want to achieve as never used unraid before. Thanks.

I am not expecting it now, however I am expecting ~4 quite reasonably (as I already have that, I just am currently having to encode multiple copies of things so users don't get fed the high quality version). And more and more content is being released in 4K HDR. Plus I need to have some side capacity for other things on the server. So it's just not quite as cut and dry as am I using that many transcodes right at this very instant. If I went with Intel, i'd be forced to do a motherboard/cpu change if I wanted more capacity (and have to wait for the XE chips to be released). I wouldn't be able to drop in a GPU as a quick fix as i'd be using the pcie3 16x slot for an NVMe expansion card and require the full bandwidth for drives. Because Intel is so behind and not supporting pcie4 fully yet the support for multiple NVMe drives is terrible, along with fairly average pcie slot throughput (the moment I put something in the second slot on most of their boards it will drop to pcie3 x8 x8, which is barely enough for two decent NVMe drives, let alone the 4 I want to run in the expansion module). So yeah, was hoping for something a little more future proof is all. Might just be a case of eating up the costs of an upgrade down the line. Just difficult as it would likely mean a new install or at least a LOT of changes to make things work. And likely wouldn't actually have release quality unraid support for quite some time too.

Hrm that is a little unfortunate. Was hoping for a bit more power than that. Unfortunately even the 10900 uses the same iGPU. That would likely be the chip I pick up, as Plex is a secondary situation for the server. The convenience of it all is right there though. Probably be an easier decision if there was easier to find stock, but hard to juggle wants via needs when there’s so few things left to buy!

Was planning to do a Ryzen build for a few reasons, primarily wanting pcie4 flexibility and because my main rig is also Ryzen so it makes trouble shooting much easier with spare parts etc. With that said, the Intel 10900 are really pushing down to a very compelling price. And I can (mostly) work around the rather crappy z490 mobos pcie4 limitations (for example I can run the NVMe drives I want off the main pcie3x16 slot via an expansion card if I don't need a GPU). My main concern, is anyone actually running 4K HDR to SDR transcoding on Plex? I know the feature is a recent addition, but i'm wondering how badly it hammers the GPU. I can't find any info on how many streams is ossible or the required VRAM on Nvidia cards per stream etc. Would love any info people can provide on it. It would be great if I didn't have to store two copies of HDR content... I think i'm still leaning towards an X570 setup and dedicated GPU, though I must admit the simplicity, less power, and likely more support for an Intel inbuilt gpu is certainly compelling... Especially with current GPU pricing.