Electric

Your post lacks insight into the following points previously addressed in my post #16, above: 1) The files from a single share can be scattered across however many drives that are assigned to it. Reconstructing the share in the absence of a map for which file belongs to which share, requires a file consolidation effort of some kind. Your phrase "Easily be read" doesn't seem to reflect this effort. Maybe the directory structure on each drive makes it easy. I have never tried, and don't want to try. 2) You can't remove a drive with failures from a redundant array and expect to recover all the data to the last bit. In the most common case, you need information from all the other drives, and the parity drive, to reconstruct the partly-failed drive. Or you are into a very expensive clean room data recovery effort that can involve replacing drive electronics or read/write heads.

Thanks. I think it is a very simple form of a Hamming code. This is the definitive work on those if you are interested: https://www.amazon.ca/Coding-Information-Richard-Hamming-1980-02-23/dp/B01N2XUF94/ref=sr_1_6?s=books&ie=UTF8&qid=1484567853&sr=1-6&keywords=richard+hamming

I ran the diagnostics script and it seemed to have completed successfully. I must have done it with the original USB stick that I have since overwritten when I restored UnRAID 5. Because there are no files or directories on the unlicensed stick that fit what you are describing. Michaels-MacBook-Pro:UNRAID $ find . -name *.zip Michaels-MacBook-Pro:UNRAID $ find . -name log Michaels-MacBook-Pro:UNRAID $ find . -name logs Michaels-MacBook-Pro:UNRAID $ That is good news.

Where would it be stored? I can't seem to find it. Michaels-MacBook-Pro:UNRAID $ find . -name *.zip Michaels-MacBook-Pro:UNRAID $

What you seem to be saying is that I would have to reconstruct the share by consolidating files from up to seven disks. Even if that's possible, it sounds like a hassle and not anything I want to get into. That's only true if there are no errors on the disk. If the Hamming codes (or similar codes) are not present, the data cannot ever be recovered if the disk has a bad sector. And the odds of having a bad sector in a 24-bay box like I have, with drives that are not exactly brand new, are not ones I'd like to bet against. Again, the risk is not zero.

Already tried it. Didn't seem to help. Thanks though. I am grateful for all suggestions.

I am not sure I will be able to read that filesystem without a copy of UnRAID that has a working web UI. The data seem to be stored across several drives in a very proprietary way. Even if I can mount the drives on a regular Linux box (can I?), I'd lose the correction codes. So the risk is not zero. It was very exciting for me when I reinstalled UnRAID 5 onto that USB key to see if I could read the data again. The UnRAID box IS the backup for all the computers in the house. I am hoping that the combined probability of a desktop/laptop drive failing, and the UnRAID storage also failing at the same time, is incredible. If I couldn't get the UnRAID 5 box working again, I would have burned another weekend backing everything up onto some other backup system like FreeNAS that I don't have time to learn right now. To answer the question about the hardware, it is a 64-bit AMD CPU in a 24-bay SuperMicro case. It has 8 GB of non-ECC RAM. It has about seven drives in it right now, including a 256 GB SSD cache drive. It has been running UnRAID 5 more or less since that version came out. Has been rock-stable; I spend almost no time administering it. I think UnRAID is fantastic.

Thanks Frank. Read it several times but it does not seem to help with the particular problem I am having. Are you aware of anything specific in that document that covers everything working except the web UI? Thanks for posting any section I might have missed. I can appreciate that many people have had a flawless upgrade experience. But I didn't ask for these problems; I can think of much better ways to spend my weekend.

Thanks for the suggestions. I loaded unRAID 6 onto a new USB key. It booted and I could telnet into it. I could not access the web interface, however. I ran the diagnostics script and it completed successfully; if I can figure out how to upload it here I will. I tried ftp but could not start a session. It does not seem to be stored on the USB key itself. I have reloaded unRAID 5 onto my registered USB key. I am now doing a parity check which will take about a day, in an effort to make sure that everything is in order for the eventual migration to unRAID 6. I am going to buy the largest external drive I can find, and back up as much important data as I can, as the migration to unRAID 6 does not seem to be straightforward for me.

I don't think so. My go file is 71 bytes, per upgrade instructions. I updated post #3 in this thread with my config files, in case anyone can see anything wrong with them. I very much appreciate your help, by the way.

When I boot in safe mode, it appears to hang at rcsyslogd initialization. See attached screenshot. Configuration files are in the attached zip files. Archive.zip shares.zip

I upgraded my 5.0.6 install to 6.2.4 (latest stable version). I did not go through any intermediate releases. I have a Pro license. The system boots and I can telnet into it. I cannot access the web administration interface. Have tried both Safari and Firefox on my Mac. Can anyone recommend next steps?

It is whatever Unraid Server Pro 5.0.6 defaults to. The disk it is on is part of the array, if that is what you are asking. In any case, I ended up creating a 50 Gig file full of zeros using dd. Then I put an ext3 filesystem in the file. How would I export that file through NFS? I have tried mounting it in fstab, and exporting it via NFS in /etc/exports, but somehow fstab gets overwritten when I reboot.

Should that command not make a file 50 gig in size?

What am I doing wrong? File size is zero. Prior to this I stopped the array, unmounted everything, then created a mount point and mounted the drive under /mnt/disk2. root@tower:/mnt/disk2/cameraBasementStairs# fallocate -l 50G ./cameraBasementStairs.img root@tower:/mnt/disk2/cameraBasementStairs# ls -lh total 0 -rw-r--r-- 1 root root 0 2015-08-15 11:36 cameraBasementStairs.img root@tower:/mnt/disk2/cameraBasementStairs# stat cameraBasementStairs.img File: `cameraBasementStairs.img' Size: 0 Blocks: 0 IO Block: 4096 regular empty file Device: 861h/2145d Inode: 1530 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2015-08-15 11:36:35.000000000 -0400 Modify: 2015-08-15 11:36:35.000000000 -0400 Change: 2015-08-15 11:36:35.000000000 -0400 root@tower:/mnt/disk2/cameraBasementStairs#

Is this likely to work on unRAID? ## Create an image file to your preferred size fallocate -l 50G /data/security/patio.img ## Format the image file as ext3 mkfs -t ext3 -q /data/security/patio.img -F ##Add the NFS mount to your FSTAB to mount the image at your NFS location (/etc/FSTAB) /data/security/patio.img /data/security/patio ext3 rw,loop,usrquota,grpquota 0 0

Nope and nope. Share has to appear to be 200 GB in size, not free space. Thanks, though.

Are you saying it is impossible even from the command line? This has been a feature of *nix since before the Carter presidency.

I need to limit an nfs share for a particular user to 200 GB. I created a dedicated username and was trying to find a way to set a quota, but failed in my effort. Using unRAID server pro 5.0.6 right now. The reason I need the quota is a firmware bug in my new Hikvision IP camera. It is allergic to shares over 200 GB. Could also use SMB, if that is easier.

Where can I get a hot swap tray for a 2.5" SSD drive, for the TAMS 24 bay supermicro chassis that I bought last May? I have tried 2 types of 2.5"->3.5" adapters from the local computer store; neither is designed for the position of the SATA connector on the SuperMicro chassis. I believe that when I got the chassis, there was an internal mount for a 2.5" drive but I completely gutted the case to replace the power supply and fans.

The fans on my AMD motherboard were the 3 pin type. I replaced the fan wall with three ziptied Arctic Cooling 140mm fans as suggested, and the 80mm fans as well. It is much quieter at the expense of a drive temp rise from low to mid 30s.

That is a great idea. Should the 140mm fans be the PWM type?

Yes, it is an AMD-based TAMS unit. I see that all the fans are 80mm and look easy to replace. They flow a LOT of air and I am a bit concerned that replacing them with low-noise fans will result in lower airflow as you say, but I figure if I can keep the drives a little over 30 degrees most of the time they should be fine. I will be installing an HX850 power supply as well.

Just a quick note to thank everyone for the support. I ended up transplanting the drives into the new Supermicro chassis. I got it to boot with the beta13 stick (which I took an image of before I moved it), but the web front end would not come up although I could get unmenu to work (starting it from a telnet session). I upgraded the stick to 5.0rc16c, simply by replacing the two bz files in the root directory, and then I could get into the web interface. I noted that all the drives were correctly identified although they ended up split on two SATA controllers (that is a fantastic feature!). I started the array and held my breath... all the files came back and it is running like a champ. At some point I will buy a set of 4 TB WD Red drives, but for now things are working great. The biggest problem I have is that the fans in the new chassis are deafening, but that is a project for another day.

Thanks to everyone for the advice. Gary, did you mean 4 TB disks? 3 TB seems to be more of a sweet spot price-wise, and I have 24 bays in this new chassis so physical space is not an issue. My current array is 4 * 1TB, and that is enough to back up my three Mac laptops using Time Machine, plus a limited number of TV shows. I am currently looking at Western Digital NAS-oriented 3 TB drives. I also wanted to ask if anyone is backing up their UnRAID box to Amazon S3 Glacier or similar? I am not totally clear on how much this would cost me, or if it is even possible.