Hi,

Woke up this morning to an unresponsive server and after a reset I plugged it to a screen to see that it was blocked on an error about a corrupted bzfirmware file.

It appears the firmware main file got corrupted as well as some other files on my USB and I looked everywhere and I don't have a backup sadly.

I managed to backup some of the file on the USB before it stopped working for good.

It gives me a "Data error (cyclic redundancy check)" error and I can't open it anymore.

Those are the files that were corrupted. I couldn't open them or copy them, it gave me a read-only error.
 

bzfirmware
/config/docker.cfg
/config/parity-check.log
/config/secrets.tbd
/config/super.dat
/shares/Main.cfg

Which from what I found appear to be all the important one 😕

So I got a few questions here.

1- What is `Secrets.tbd` and how screwed I am from losing that.

2- There is a 6 months old `tower-diagnostics-20231120-2001.zip` file that contains my `Main.cfg`, `docker.cfg` and `super.dat` files. Is the `super.dat` file the real thing.. or has part of the info been retracted for privacy reasons, like how the `Main.cfg` file has is name partially masked by dashes?

3- Any suggestion of a software that I could try to rescue the files or is it fried.

I only have one parity drive so from what I got it should be easy to restructure everything. The parity drive should be the one that I can't mount.

Any suggestion on how to proceed with that? I am not much of an unraid expert. Just bough a new USB drive so I can reinstall unraid on it.

Thanks!

Edited May 18, 20242 yr by devilwarriors
Typo

The USB drive would no longer mount on windows or mac, but I tried it on linux on my stream deck and I'm able to mount it.

The same files were corrupted, but somehow `super.dat` copied fine. Not sure if it's uncorrupted yet, but some hope there.

So now with this I'm only missing `secrets.tbd` and `parity-check.log`, assuming I can use `docker.cfg` and `Main.cfg` from the log zip.

I tried running fsck.ext4 but it complained about the drive being read-only. I ran it with the `-n` param and it told me that I have a bad magic number in the super-block...

Edited May 18, 20242 yr by devilwarriors

if you can read the content from a linux envrioment you will need to copy the entire config folder and get a new usb and copy that config folder to the new unraid USB.

Otherwise its data loss and non-recoverable...

It is important to download a backup flash drive when in unraid. plugins and tools exist to automate this to once a month etc...

8 hours ago, bmartino1 said:

Otherwise its data loss and non-recoverable...

Note this is only referring to the USB drive, not the Unraid controlled drives.

So I got lucky and the Frankenstein of a USB drive that I rebuild form the files I managed to pull from the old drive and diagnostic files seem to have worked.

I did had to reselect my drive, so maybe super.dat was corrupted, but everything is back and the dockers even work as normal without me doing anything.

Would still love to know what I lost with losing `/config/secrets.tbd`.

But that server is decided to ruin my week-end and so I woke up this morning to my largest and most recent disk having been disabled and now being emulated.

Last night it threw 2 smart errors for the first time before bed.

Here is the smart file, any idea if it's dead already or what?

ST12000VN0008-2PH103_ZL2EM5E5-20240519-0728.txt

super.dat is the hd file and is used with the array and pool. super.dat is essential the HD "superblock" where data is located and accessed if encrypted.

the secrets.tbd is samba/unraid account passwords and is used with samba and sometimes ssh:
https://wiki.samba.org/index.php/Frequently_Asked_Questions

Glad you have recovered

Please go to main ? Flash > download flash backup! and store it on IDK google drive or some place safe to return to a good condition in the event this happens again.

please go to tools diagnostic and post the full diagnostic. We can then assist on next steps to reconnect drives. If they are throwing smart errors it may be caused by the superdat file being corrupted and it may need to be rebuilt... otherwise since you have smart errors you may need to replace the drive soon.

in essence you need to power off. delete the super .dat file and let unraid recreate it.

you drive configuration should still be there and in good order just re-select the disk and let unraid run parity checks.

Concerned as you have a drive with smart errors that can corrupt parity / the superdat file and should be replaced sooner than latter. Or power off to save the data on disk for other recover, such as dd to another HD.

Here the diagnostic.

The error the drive give is "UDMA CRC error count". I've read on this forum from a google search that it's possible it's the cable that is badly connected and I did open the back of the case which is quite tight. It's quite possible I pinched a cable or it's not connected completely. Is that a possibility from what you see from the smart file or does it really look that bad?

Is it safe to flush the super.dat file while one drive is down or should I empty and remove it first?

I saw a video tutorial on how to use the unbalance plugin to empty a drive, do I need to do that or can I just remove it from the array and it will put the emulated data on the other drives?

I'm lucky that it currently only has 2.5Tb of data that could be moved easily on the 3 other smaller drives.

tower-diagnostics-20240519-1738.zip

Replace cables for disk1 and post new diags after array start.

UDMA CRC error count

in a hba

if you don't follow 1,2,3 plkug order this will go up.
Ther is a bend or break in the sata cable that cases this # to go up when read/written too.

It is recommended to replace the data sata cable.

UDMA CRC error count is also a sing of old age.
https://www.minitool.com/partition-disk/udma-crc-error-count.html

Quote

in a hba

if you don't follow 1,2,3 plkug order this will go up.

Can you clarify what you mean here? I think your talking about the order that the sata cable are connected to the motherboard.. but a quick google search lead me back to a post on this forum saying it doesn't matter.

The disk is pretty new so I hope it's the cable, I will replace it as soon I find one and will report back.

Thanks

Quote

Replace cables for disk1 and post new diags after array start.

Done, Here the diagnostic.

tower-diagnostics-20240522-2222.zip

Emulated disk1 is mounting but it's empty, did you format or was it already empty?

I've ran the unbalance plugin just to move the stuff to another disk but haven't went ahead with zeroing it or any of the other step that were talked in that video.
 

Awaiting further instructions before I do anything else.

Would for sure rather save the disk as it's pretty new.

Not sure I follow, did you move the data that was on disk1 to another disk(s)?

Yes I moved it to disk 2 while the disk was still emulated in case I need to get rid of it.

OK, in that case you can rebuild disk1, you can then move any data back there, if you want.

Quote

Is it safe to flush the super.dat file while one drive is down or should I empty and remove it first?

I had that question earlier, should I still go ahead with a reset of super.dat and should I do it after or before rebuilding the disk?

Not sure I follow, you usually don't touch that file, just rebuild the disk normally.

Quote

please go to tools diagnostic and post the full diagnostic. We can then assist on next steps to reconnect drives. If they are throwing smart errors it may be caused by the superdat file being corrupted and it may need to be rebuilt... otherwise since you have smart errors you may need to replace the drive soon.

in essence you need to power off. delete the super .dat file and let unraid recreate it.

It's @bmartino1 who mentioned this a few messages up.

I'll just rebuild the disk normally.

Edited May 24, 20242 yr by devilwarriors
Phrasing

Lot of confusing information about how to rebuild a disk.

Do I stop the array, remove the disk, run a preclear on it and then put it back and it should rebuild?

Any tutorial you could link if it's not that?

Thanks!

Edited May 24, 20242 yr by devilwarriors
a word

If you want to rebuild a drive to itself this is covered here in the online documentation accessible via the ‘Manual’ link at the bottom of the GUI or the DOCS link at the top of each forum page.

 

Running Preclear on a drive prior to a rebuild achieves nothing as the rebuild process writes to every sector on the drive so existing contents are irrelevant.