Hi

 

I had a red-ball a couple of days ago. I have about 14 drives and many are 3+ years old so this wasn't particularly unexpected, although it was a little disappointing that the drive in question was a newer 2TB one (maybe 1.5 years) rather than one of the older ones. Anyway, I bought a new drive (WD 3TB) and swapped it in. But instantly upon starting the data-rebuild it red-balled again.

 

D'oh! Probably a cable, and perhaps I've moved too quickly to swap out the disk.

 

I've replaced the cable now and I can smartctl on the disk fine (it shows zero on pretty much everything but spin count and temperature, so I'm pretty sure the disk is ok!) But now the unRAID slot appears to know it's supposed to be 3TB, not the 2TB that it was a few minutes before. So suddenly I'm not sure how to make it recognise the old 2TB properly, nor to have it try again to data-rebuild onto the red-balled but (probably) fine new blank 3TB.

 

So, I have a couple of questions that I was hoping someone could help me with.

 

1 - Is there any way I can reasonably easily prove that it's a cable (and not say a controller problem that only happens when I try to write lots of data during a data-rebuild). I thought of doing a pre-clear on the new 3TB, but it thinks it is still part of the array so I imagine pre-clear won't want to touch it right now.

 

2 - Should I push forward and try to get unRAID to do the data-rebuild onto the 3TB somehow even though it has already red-balled that disk; or should I try to get the old 2TB disk back where it was and start again with the 3TB after that dust has settled?

 

3 - Is there any way I can do some of the above steps without a parity rebuild (I'm not super strongly attached to the parity, and if it needs to be rebuilt during/after my recover then that's fine... but in a perfect world I'd find a solution where it'd be like I went back in time and just replaced the cable before it even failed, and now happen to have a new 3TB disk to add somewhere!)

 

My theory (which I thought I'd check it through with the community first this time before rushing in!) is that I can pull the 3TB out and put the 2TB back where it was, then unassign it, start, stop, re-assign, rebuild parity, and I'll be back where I was. Then the 3TB is just a new disk to do whatever I want with. What I don't like about that plan is that I'm still mildly suspicious that something other than a bad cable is at play here, and if I discard my parity and I just happen to have a broken old 2TB and a broken new 3TB then perhaps that'd be bad...?

 

Anyway, thanks for your input. I have rarely had issues with unRAID but ifI do, you lot are always there for me!

 

BTW I'm running 5.0rc5, which I know is a bit old but it probably shouldn't matter. I'm mainly interested in the procedures involved with my situation. unRAID seems to be doing everything right, I'm just not entirely sure what the best steps are to get from where I am to where I want to be.

 

-- tallorder

If you're CERTAIN that the old 2TB drive is good [i.e. have you attached it to another PC and confirmed all the data is readable?] ... and truly don't care about doing a rebuild based on the parity information ... then you can simply do a "New Config" and assign the drives you want to that configuration;  then let it build parity from scratch.

 

If you do that, however, be SURE that you assign the correct parity disk (safest approach is to not assign parity at all until you've got all the data drives assigned and confirmed).    And of course if you're wrong about the 2TB drive being okay, you'll lose all the data on that drive.

 

Remember, the whole idea of fault-tolerance is you can rebuild the data from a failed disk onto a new one.  So be sure you don't need to do that before you reset the configuration.

 

Hi

 

Thanks for the advice. I'm not feeling especially cavalier about  or rebuilding parity, and will only do so under circumstances where it is relatively safe to do so (i.e. I hope not to have to do it at all!) On the plus side the majority of important files are backed up via CrashPlan, so I'm not crippled with anxiety right now either. But I am not yet sure about the old 2TB drive, and won't be able to test it for 24hrs or so. For the purposes of this thread lets assume that it is toast, maybe physically failed along with the cable or simply corrupted somehow.

 

So my updated problem statement is that I have replaced a failed disk with a brand new 3TB that I'm pretty sure is fine, but which unRAID has red-balled due to a bad cable. I can't discard my parity. I'd like to get unRAID to retry the data-rebuild onto this disk, and I have some hope that it will succeed this time now that I've replaced the suspect cable. How can I get it to try again?

 

Thanks,

-- tallorder

Stop the array and power down;

 

Re-seat the drive and its cables (perhaps replacing the data cable); 

 

Reboot; stop the array;  un-assign the drive;  Start the array (it will show a missing drive);  Stop the array and re-assign the drive to it;  Start the array again ... it should now rebuild the drive.

 

 

Brilliant, followed your instructions and it's behaving exactly as I hoped it would. Rebuild underway!

 

Now, time to test the supposedly failed old 2TB and see if it's ok. I suspect it will be the bad cable that's been the problem all along and the drive was never failing. Oh well, any excuse to expand the array, I guess

 

Once again, thanks for the help!

 

-- tallorder