Using chkdsk on one drive preventing farming

Hi, this morning I decided to run chkdsk on one of my messed up drives.
As soon as I started the chkdsk, my chia stopped farming. It was still synced with the network, but no plots were going through filter.

WHen windows runs chkdsk it locks the drive from being accessed. However, this seems like a chia issue, because when I went to try and add/remove drives from my plot directories (in the GUI), chia gave me an error message saying that I cant add/remove drives because one of the drives in my plot directories list may have a problem i.e. the drive which is being checked.

So when one single drive in the list is locked, then chia cannot access any of the drives for some reason. So I just removed the drive from my config.yaml and restart, then chia works again !

If you run chkdsk in read-only mode (i.e. don’t specify the /f switch) I have never had an issue with any drives containing plots whilst Chia harvesters and/or full node are running.

If you force a volume to dismount (it will prompt you to do this is not if read-only mode, and the drive is already locked) then I am not surprised Chia has issues accessing those plots - because the OS has dismounted the volume.

You can always try an online repair if there are issues (/scan switch), otherwise the only other option if repairs are needed is to schedule the fix for the next boot.

I understand that chia cant find the plots on that drive, but I don’t understand why chia couldn’t access ALL my other drives until I removed the messed up drive from my plot directories list.
That is the confusing part…
One inaccessible drive messed up my entire farm haha

Could be a Chia issue with handling dead drives, could also be OS not handling flakey drive as well as it could and passing the problem to apps using that drive. Drives can fail in all kinds of wierd manners from my experience, some quite clear cut and some producing all manner of strange behaviour depending on electronics/hardware failure modes.

How full was the fail(ed/ing) drive ? :laughing:

Well I did a full format before chkdsk so 0% full haha.

Wanted to post here in case someone else experiences the issue or knew of why.

Maybe its just my OS like you said :confused:

Did you do a ‘quick format’ ? This is ok for new drives, but I always untick that box (or use command line) to do a full format.

Quick format just writes the FAT and sets up journal and a few other checks etc. A full format actually scans the entire drive looking for bad sectors and marks them accordingly.

chkdsk won’t check for bad sectors either unless you tell it to. It’ll just cross check the FAT/journal and make sure both match with the actual files on the disk (not the content of the files).

Can you share the exact command you used? That’s interesting how all the drives were locked when only one should have been.

1 Like

I’ve had this before, but a different scenario, and I put it down to poor programming in Chia.

When I plot a drive on my workstation I share that drive and add it to Chia, but if I forget to remove it from Chia prior to taking it offline then I get the same problem - Chia will not let me delete it, its says there is a problem with one of the drives in my plot directories list. I have to close Chia, manually edit the config file and start Chia again.

1 Like

I only said one drive is locked, the drive on which chkdsk is being run. Which is normal behavior (It is dismounted)
BUT when that one single drive is locked by chkdsk, then chia does not know how to handle the scenario and it stops farming. All other disks can be accessed normally in Windows through file explorer. There are no problems there
It seems like @Ronski has experienced something similar, which also tells me that both our conclusions are accurate… a problem with chia :smiley:

1 Like

I’m sorry, my mistake.

1 Like