I have a system with dual parity running v6.2.4.I have messed up and need to do a similar "Re: Forcing Unraid to recognise an older disk configuration".

My attempt was to swap or migrate 2TB drives from a 2 port Vantec SATA card to a 8 port Adaptec 6805 card.

My test of the Adaptec card was by adding a new 3TB drive with (NTFS) data on it.

I mistakenly chose the wrong (which I thought was a spare)2TB drive and added it to the Adaptec card in the default "array mode".

The Adaptec card presented it (the 3Tb drive) as a "new array" option in the Adaptec bios and not as JBOD mode.I know that now as I see I have to initialise any drive added to the Adaptec 6805 to get it in JBOD mode for unRAID to see it which means losing data.

Changes were made in the Adaptec bios and I went ahead and booted the Tower.

Attempted a new config to start fresh to add more drives.

I have added the Vantec card back into the system in order to go back to the original config in an attempt to recover data on the lost drive(s).

I have not started the array with the new config.

My Tower does not start the array on boot.

Can I recover data on any data drive as I have dual parity?

Is the new config revert an option?

Thanks

towering-diagnostics-20050118-0123.zip

I am asking if I put the first parity disk with all the data and cache disks back to their original slots and select trust parity option before starting the array,it should discover missing data disks and rebuild? yes/no ?

I am asking if I put the first parity disk with all the data and cache disks back to their original slots and select trust parity option before starting the array,it should discover missing data disks and rebuild? yes/no ?

No

 

Your initial post isn't very clear, and neither is this one, at least to me.

 

If you put all the disks back to their original slots and select trust parity that may be a good starting point for deciding what, if anything else, needs to be or can be done. But I don't understand why you say you will have missing disks if you put all the disks back. Whether or not you need to or indeed can rebuild any disks remains to be seen.

 

Thanks trurl.

Your initial post isn't very clear,

The new config rollback option is not available is what I am asking as I have trashed one or two disks using the adaptec card in array mode.

But I don't understand why you say you will have missing disks

Not missing- deleted data

if you put all the disks back

I have a snapshot of the disk slot positions

or indeed can rebuild any disks remains to be seen

Original config and starting the array with "trusted" parity

 

If you have a single parity disk then it can only rebuild a single data disk, and only if all other data disks are correct.

 

Do you have backups?

If you have a single parity disk

I have dual parity.

and only if all other data disks are correct

I am assuming correct as in slot position or having done a parity check recently?

Do you have backups?

I have older copies of data.

If you have dual parity, then you can rebuild 2 disks if all other disks are correct. Correct as in correct position, and also correct as in every bit of the contents of those disks are correct and can be read so the missing disks data can be calculated. Parity doesn't contain any data, it just contains bits that can be used to calculate the data of a missing disk from the data on all the other disks.

The new config / trust parity method is useful if you want to put all of the disks back into an array - and you know that all of the disks are accurate.

 

It sounds to me like you want to use this method when you know that the data on one of the disks is faulty (i.e., has been deleted). You want to put all of the disks back into the array, including the one with deleted data, but you want the one with deleted data to appear dropped from the array. If this were possible, you feel you could rebuild that disk.

 

Is that what you are asking / trying to do?

How many drives do you think may have been affected by whatever you did?

f you want to put all of the disks back into an array - and you know that all of the disks are accurate.

Yes bjp999 .The disks are still in their original positions.

It sounds to me like you want to use this method when you know that the data on one of the disks is faulty (i.e., has been deleted)

Yes I selected a used instead of removed array disk on the Adaptec card.

but you want the one with deleted data to appear dropped from the array. If this were possible, you feel you could rebuild that disk.

Yes.

Could a rebuild be possible of these "incorrectly configured Adaptec migrated" disks?(only 2)

trurl

How many drives do you think may have been affected by whatever you did?

I had a 2 port Vantec SATA card running the affected disks(2TB)

If all other disks are intact and you know their order you can rebuild them.

 

Do a new config, assign all disks, they all need to be in the previous slots, same order, then and very important, check the box "parity is already valid" before starting the array, start array  (ignore if any of those two  disks is unmountable), stop array, unassigned both disks you want to rebuild, start array, check that both emulated disks mount and data looks OK, if everything looks good stop array, reassign both disks and start array to begin rebuild.

johnnie.black

Do a new config

did that but did not assign any drives

assign all disks

need to be in the previous slots and same order

check the box "parity is already valid" before starting the array

start array

ignore if any of those two  disks is unmountable

stop array

unassign both disks you want to rebuild

start array

check that both emulated disks mount

check that data looks OK

if everything looks good stop array

reassign both disks

start array to begin rebuild

thank you.I will confirm results

I have a slight concern that when the array comes online, that a small amount of writing will take place while the 2 disks with errant data are still in place. Would not want turbo mode. Would be nice to take them offline before starting array - but not sure unRAID would allow that.

johnnie.black

I did as prescribed and I have my data back.On my first array start I noticed all drives were marked green.

In emulated mode I could see disk 4,5 disk data.

johnnie.black is your

check the box "parity is already valid" before starting the array

option supposed to be checked only for the first stop and start of the array?

I have attached a pic of the current state.Is it normal to have so many parity errors now?

bjp999

I have a slight concern that when the array comes online, that a small amount of writing will take place while the 2 disks with errant data are still in place

Which stage in the process were you referring to?

Would be nice to take them offline before starting array

How would the disks be read offline?

Would that come as a new feature or request to LT or in the current unassigned devices tab?

johnnie.black is your check the box "parity is already valid" before starting the array option supposed to be checked only for the first stop and start of the array?

 

Only after the new config (it's also the only time that checkbox is available).

 

I have attached a pic of the current state.Is it normal to have so many parity errors now?

 

No, and if they happened during the rebuild both rebuilt disks will have at least some corruption.

 

Don't reboot and post the diagnostics: Tools -> Diagnostics.

johnnie.black

I have attached the diagnostics.

Thanks

 

towering-diagnostics-20050118-1715.zip

Errors started soon after the rebuild:

 

Jan 18 08:41:33 Towering kernel: md: recovery thread: recon D4 D5 ...
...
Jan 18 08:55:47 Towering kernel: md: disk0 read error, sector=152026632
Jan 18 08:55:47 Towering kernel: md: recovery thread: multiple disk errors, sector=152026632
Jan 18 08:55:47 Towering kernel: md: disk0 read error, sector=152026640
Jan 18 08:55:47 Towering kernel: md: recovery thread: multiple disk errors, sector=152026640
Jan 18 08:55:47 Towering kernel: md: disk0 read error, sector=152026648
Jan 18 08:55:47 Towering kernel: md: recovery thread: multiple disk errors, sector=152026648

 

SMART for parity looks mostly fine, probably a cable problem, I would power down, replace both SATA and power cable on the parity disk, power back up and repeat the new config procedure.

 

johnnie.black

I did as prescribed and I have my data back.On my first array start I noticed all drives were marked green.

In emulated mode I could see disk 4,5 disk data.

johnnie.black is your

check the box "parity is already valid" before starting the array

option supposed to be checked only for the first stop and start of the array?

I have attached a pic of the current state.Is it normal to have so many parity errors now?

Give us a screenshot of Array Operation. That is where you will see parity sync errors.

 

I'm not absolutely sure about the parity disk, but for all other disks the errors column on Array Devices is read/write errors on the disk.

trurl

Shutdown down the server to replace data and power on parity one

Attached main and array operation pics

Repeat the complete procedure, like yesterday:

 

Do a new config, assign all disks, they all need to be in the previous slots, same order, then and very important, check the box "parity is already valid" before starting the array, start array  (ignore if any of those two  disks is unmountable), stop array, unassigned both disks you want to rebuild, start array, check that both emulated disks mount and data looks OK, if everything looks good stop array, reassign both disks and start array to begin rebuild.

Thank you johnnie.black.replacing sata cables and will continue with procedure.

 

johnnie.black

Current state attached - after powering on.

Ran a SMART short self-test: no issues

Do I proceed as instructed?

Thanks

Yes, proceed with the new config.