cammelspit

I do have a backup but not recent enough. It hadn't even been long enough for my monthly flash backup to kick in. Yes, I do boot from a USB2 port and rebooting the server works, except this stupid GUI thing. Is there a way to know what is actually corrupted file system wise and replace just those files? I thought it would last longer than 30 days being new as my old flash drive lasted over two years and it was already 5+ years old and heavily used. This is all incredibly frustrating. I'm moving in just a couple of days so I guess ill just redo the drive from scratch AGAIN once I get all settled in the new place, unless there are some suggestions on fixing it as is.

That would be silly... I mean I JUST had a flash drive failure less than a month ago. Brand new never before used flash drive with a brand new install of unraid. Anyways, thanks for helping me try and figure this out, and I was able to tab around the UI enough to get the diagnostics. If it is the USB drive AGAIN, what is the literal most durable highest quality USB drive on the planet? connollyserver-diagnostics-20200107-1903.zip

Most of the GUI is corrupted with this, this is new as of just a few hours ago and I have no idea what it means. Any help would be appreciated.

Cache size is 1TB. Global settings is set to 200GB, I don't remember setting this manually. The cache at the time had about 500GB or so left on it, I ran mover and now it is about 600GB. I dont know, maybe I am getting these writes from somewhere else? Running smbstats showed only my username I use for my own personal access from my main PC and the nVidia Shield I have that has mounted the server shares using it's own user name. just those two IPs. Thing is, I am running a good deal of docker images right now but some that access often should be running off shares that are set to 'prefer' or 'only'. I am sure to do some more digging but I have to say, I turned off docker entirely and rebooted the server without docker running at all and I was still seeing writes tot he array directly. No idea what could be doing this. Thanks for clarifying the 'prefer' option's function. Ive had phantom network access a few times that needed to be tracked down but never phantom writes to the array that are obviously ignoring my settings.

Is it just me or is the Prefer setting writing tot he array first and mover is moving the files to the cache drive, like the reverse of the 'yes' option? I thought this was supposed to be the same as 'only' except that it would write to the array if the cache is full... Am I missing something here? This is the first time I have tried to use this setting so I would like to have some clarification if possible. If this is how it is supposed to be working than it's basically a useless setting, right? Because the description seems to say that it should still be acting like 'only' except for being able to use the array in case of low disk space but it sure as heck is writing to the array first and then mover is moving to the cache disk. The big issue is that I have lots of little baby writes going on all the time, which is what the cache is there for. Maybe I am doing something wrong? Below is basically how I have all my shares set up right now that I want to use the cache. For a little bit of background, i just had a USB drive failure so I went and upgraded right to 6.8 but I have been running the older version from my sig since the moment it was release till just a couple days ago.

Damn fine job on this. Only thing I am really missing over the manual YT-DL over CLI route is being able to easily get the thumbnails, captions and other metadata. Otherwise this is a nice, and more importantly, easy way to do this. I am eagerly awaiting how this app shapes up in the future for those couple niggling prerequisites I need to toss this in full production mode on my server. Thanks a heap for taking the time to do this. Would be epic to be able to scan the directory and make sure each video is present and to fill in the gaps even if they are randomly missing. Friend of mine made a python script that does that sort of thing. Maybe a list of the videos on a channel and the ability to skip? I know these are more wish list items for long in the future but I thought I would add my two cents as to how awesome I can see this thing getting. Again, thanks, and good work!

HI! I would like to keep a specific file type on the cache drive during a move operation. long story short, I am downloading a torrent that is pretty large, in fact, it is larger than my cache drive is. I have the torrent client set up to download to a directory that is only of the cache drive first and then once it is finished it will move over to a share that the mover will work on and then it is moved over tot he array. Basically, since I can't fit everything for that one torrent on my SSD cache drive at once, this method doesn't work. So I am looking for a way to keep the partially downloaded files on the cache but to allow the already completed files in this or any torrent to still move even if it has not fully finished, by this I mean *.!qB files. If I can do this, I may just stop using the 'move to this other share so the mover can move it' method I have had in place for so long. Currently I am having to manually move the completed files over to the array and shut off the download while I am manually moving stuff, make 100% sure the file structure is correct, hope I caught all the already finished files, and restart the download. Seeding off the array is fine for me, of course, but downloading to the array partials is 100% a no no for me. I want to keep all the files I copy over full, intact, order, and on the correct drives mover would use, which is a real pain to do manually every time I get close to filling up the cache. Not to mention the IO hits downloading and seeding would have for my poor spinning rust. Any help would be greatly appreciated, THANKS!

I support just myself on my phone on my Plex server and I use a reverse proxy. If you have any more than one or two services, it actually becomes easier to use a reverse proxy instead of port forwarding. Then there are those services that simply do not work at all properly in many areas without one, like next cloud. It also reduces or even eliminates the need for using a VPN to remote access your server. Hoopster has a very valid point that it doesn't give you any real security concerns forwarding specific ports to specific services but it does make it more convenient from a user point of view, IMO. plex.domain.com is way easier to remember than an IP address. Personally, I have a good 10+ sub-domains proxied to services on my server and I like the convenience a lot. You also do have that peace of mind knowing all the traffic is SSL, which is nice. To be fair, I would 100% consider this opinion. Especially since I already owned a domain I have had since I was 11 years old that I was already paying for and also one for my Gsuites account. NginxProxyManager also makes it very easy to do. For me it was an easy decision to make.

I just found out about riot.im and am looking at a way to wean myself off of every other chat, voice, and video service out there that I don't run myself and this looks like it would fit the bill nicely... Alas, I just don't have the knowledge to be able to set this up myself as it seems like it is rather complicated if you don't know what to do. I would donate a minimum of $20 to anyone of the big guys who would be willing and able to to make this something I can work with in unRAID. No joke, this would be PERFECT for my needs. I'm not too noobish and I have set up nextcloud, pfSense VM, reverse proxies. Grafana, that one was a BEAR but oh so useful when you finally get it working smoothly. So I have an idea of how to set up a lot of the more complicated dockers and such, just not enough to do that part myself as I am still learning.

Have room for two more drives which would kick me barely over the 100TB mark. Maybe I can convince the wife a JBOD enclosure would be nice for Christmas this year? 😜🤞

The epic of my server seems to finally be over... The last drive of all of them finally finished preclearing this afternoon. I added to the array and added one to the second parity slot just a few minutes ago the last drive. So far, every bit of data is there and accounted for. The array is now 82TB but now also has that second parity disk, which is building right now in the background. When I sparked plex back up after switching back to my pfSense VM for internet and everything just hummed back to life like nothing was wrong, it made me so happy. It's like my server was just turned off and let sit there for a few weeks for no reason and nothing running on there is the wiser that it was a bumpy ride. Turns out that the new battery somehow bugs out if I have the USB plugged in while the server is booting, the fix is to just run it without the battery connected to the server. Of course, this is a little annoying because I won't be able to have the server automatically shut off this way in case of a power outage but I'm home almost all the time so I think the better than a half hour runtime is probably good enough for me to do it manually if something goes wrong. At least until I decide to bite the bullet and get the network card for the battery unit and get a script put together that can do it for me. This is the least of my worries anyway. This will probably be my last post about this adventure that I have been on here and I leave with a smile on my face. Sure it was expensive, sure it was very sketchy for a while there, sure it took WAY longer than I expected... BUT! Not one bit of data was lost, Not one drive failed the preclear at all or showed any errors of any kind. I'm happy. Words to the wise, always keep your backups up to date. If you have expensive hardware or data you can't afford to lose, back it up and get a good battery and save yourself the headache I had, it's worth every penny.

hehe

Just a quickie update. Last two drives have finally started to copy their data onto the 'new' array. so far every drive without fail has passed the preclear with flying colors and no read/write errors getting tot data back onto the array. Soon as this finishes copying, it will be time to switch back over to my pfSense VM and stop using Panicrouter. Yes, Panicrouter is actually the name I game to my old Linksys whatever the heck it is that I am using to keep from having the internet go down every time I take the array offline to do anything. I am losing a good 50% or more of my speed and my latency has been toilet level since I switched over, so that will be nice. At least I can build the second parity disk while the server is online and properly functional, I just won't be running any of those automated services till the second parity is built. To anyone who is interested, the new battery backup has been working great in double conversion/online mode. The thing is so much quieter than I expected. It is a Tripp Lite SU1500XLCD in case anyone was interested in which model I grabbed. I don't have a proper rack set up right now and even when/if I do, this can just sit next to it as getting the floor box version saved a few bucks. Anyways, hopefully, I won't be posting, implying I haven't had any problems, till at least after the last copy from the last two orphaned drives is done. An adventure this has been but a learning experience as well. 😜

Got all the new drives installed ina brandy new array on their own, I am now copying files from the unassigned devices plugin into the new drives until the old drives are clear and then running preclear on those freshly empty drives. I have precleared BOTH of the drives that failed before, including the one that I thought was clicking and they both not only allowed me to save every bit unharmed, but a preclear operation on them both finished without a single error. Could I have just heard a cable get into a fan somewhere? I mean, there are nine fans in my system and I have done the best job I could with cable management considering the current 14 drives crammed in the front of the server chassis. I did pull and reseat literally every SATA data and power cable in the entire machine and I can tell you at least one of those cables wasn't clicked in all the way, though it had never been a problem before. All I can do now is trust that the drives were just very confused and tagged bad by unRAID because of the power surge and that everything is probably fine. I mean, they both are less than a year old and weren't even that heavily used really, which is why saving the data off of them wasn't actually that bad. Those two drives are just waiting till I transfer everything from some more of my other 10TBs onto the new 10TBs fully so I can stop the array and add them and two of my older 4TB drives at the same time as the second parity drive. When you have almost 40TB out of a 62TB array to copy over MAYBE two drives at a time, it seems it can take quite a LONG LONG LONG time. LOL 😁 Thank the powers that be, I haven't lost a single file so far and will run file integrity soon as everything is copied over. I canceled my order on Amazon for the new battery I was going to get. It was a Line interactive unit, way better than the POS standby unit I had, but honestly, that wasn't enough for me to feel perfectly safe that power issues wouldn't end up being a problem in the future. So I hopped on NewEgg, they had a better price, and I got a big ol' Online/Double Conversion style unit. Yes, yes it was more than twice the price but the full electrical isolation from my older homes wiring was worth the investment in my opinion. The new battery is rated for 1350 Watts, max and even though it is 55 pounds and absolutely massive, it is whisper quiet and performs VERY well, better than any battery unit I have ever owned in my life. It's a real pro level unit too with expandable runtime if I ever felt that was necessary. It's the isolation and zero switch time that really sold me on it. I mean, it is to protect what has become a very large investment for me and the peace of mind alone is probably worth the price of admission anyway. Well, this won't be my last post about this of course but I won't be making too many for a while at least because it is mostly just lather, rinse, repeat, with copying files, preclearing drives, adding them back to the array. Probably will be at least another week and likely a bit more till I can spark everything back up and hopefully return to normal operation. Wish me luck!!! 👍

Upgraded my license, preclearing the new drives, finishing command-line recovery off the drive I thought was clicking and so far, nearly all the data is good and safe. No idea how it could click before yet now seems to be fine but I am not complaining right now. If I can save the data, that's all I care about. I have a partial backup to a paid for business Gsuite account. The only reason why the backup is outdated is the fact that my internet, even though I have 1Gb down, I only get 35Mb up. I have been permanently uploading for over 8 months now and it just hasn't finished. At least my nextcloud data is all 100% safe and fully backed up as well as the hardest to replace data I have. The only thing that is even partially losable is some of my more recent media. I did it in this order so that if something ever went wrong I could restore what is most important and irreplaceable. All in all, I am just happy I went with unRAID over any of the other methods, I mean, with a hardware raid or ZFS raid using something like FreeNAS I would have lost everything in one fell swoop. Sure it's gonna take me over a week, maybe longer, to get back up and running normally but I like the idea of the small win first. I know this has become more of a blog about what's happening to me but I started it and it might still be interesting to some out there.

😖 😪

And now I realize I have unRAID Plus and not Pro and you can't start the array with more drives installed, even if they aren't going into the array. Yippy! 😩 Man, my day keeps getting better by the second. I know two of these drives aren't even going back into the system, making me be under the limit of 12, except I would have to unplug two drives and I won't even be able to pre-clear all three new drives with the array started in emulation mode. I will say it would have been nice if the fact that even OUTSIDE the array you can't have more than 12 drives installed at all was stated more clearly otherwise I prolly would have just gone and bought Pro, to begin with. After spending $1000 on the new drives and a new battery backup, with 'pure sine wave', since mine is obviously not working well enough to keep surges from destroying my server, what's another $50 right? 😕 You know what? I am going to shut my server completely off. I think I have had it for tonight and I don't think I can even stand the sight of it right now. I'll get some sleep, get back at it in the morning all fresh and bright-eyed. 🤷‍♂️

I am only lucky in as much as I can afford it and my wife let me do it. She's a saint letting me toss so much cash at this. 😜

Just went out and bought 3 new 10TB drives. The one that's totally dead is an 8TB so a small upgrade. The repaired one may or may not be good and it is also an 8TB drive. If the xfs_repair drive is actually physically good and passes preclear, I will then look at replacing one or two of my older 4TB drives. I also really want to stick with having two parity drives now to keep this from happening again. Looks like my next week will be taken up by and full of preclearing drives and copying data over to a new array since this one is basically gone and I want to confirm which if any of my drives died. Though I am always open to recommendations if anyone has any.

Currently copying data from the dead/dying drive now via the command line onto other array drives. assuming I don't have any other failures, I should be able to save the lion share of the data. I just need to get a few more drives so that I can have a staging area to copy stuff off of the current drives and run a preclear to confirm that they are physically intact. It may also be time to drop in a second parity drive too. 😒

ARGH! seems like one of the drives is clunking.... and another had filesystem errors and had to run XFS_reapair the data on it was saved but now unRADI cant use it as a drive for rebuilding or emulating the dead one. Ugh....

Hey, so I know this may seem odd, but I had a power outage this morning and it caused read errors on a bunch of my drives. See, I do have a battery backup unit but it's not a particularly nice one. Anyways, I rebooted the server and all errors magically washed away only for some reason one of my drives has an X by it and it won't let me do anything with it, there was even notification about how drives no longer have errors. How do I ignore this red X and let the drive come back? I know 100% without any reservations that the hardware is fine and the data is prolly perfect still too as the server wasn't doing anything really when the power went out. The drive that is being emulated was one of the drives that had almost literally nothing on it save from a few random files here and there. I just want unraid to stop emulating the drive, that's it.

THAT worked, thanks a bunch!

It does this in chrome as well and there is nothing to copy paste, I tried. Unfortunately, it seems to be unrelated to your issue.

Hi, when I open the terminal window, it remains completely blank. I also cannot SSH into my server to use the command line that way either. I have cleared browser cache to no effect. Since a picture is worth a thousand words, I figured a video would be worth a thousand pictures. I also try in two browsers, the Firefox install is 100% fresh and I have never used it, same results. Thank you in advance for any assistance you guys can provide!