how to force a rebuild on a data drive

[Neo_x]

Hi guys

 

I had a stressful morning - with some very odd problems on my Unraid server.

 

I noticed the last few days that i have a suspect data drive( drive works 100% - but the moment i power down the system and up again, the drive have the tendency to be missing from the array, and on a 2nd or 3rd reboot (and alot of cursing / moving/swapping of cables in between )  - it returned.

 

in any event  - i completed a preclear and parity check yesterday , and all seems well - but this morning i encountered the same issue again after power down , but this time unfortunately the drive just didn't want to return to the array.

 

 

thinking that all is lost, i then moved the drive to my "trusty" windows system, from where it was detected immediately - not exactly sure how that could be, since i tried difference SATA ports on the unraid machine to no avail. i even kept the drive powered via the Unraid PSU with just a sata cable to the windows machine - so i was stumped.

 

in any event, i thought it feasible to try and clean the drive - which i did via windows disk management deleting the partition and creating a simple windows partition on it. (note - i didn't do a full format - windows probably only did a quick format..)

 

then i continued on to get a firmware upgrade from seagate(which the drive was a few versions behind). this passed without issues

 

 

all is well - my expectation then being that once i move the disk back to unraid i will need to restore the data onto the drive.

 

 

problem being - powering on my Unraid system with the drive connected back onto the array, unraid started without reporting any problems on the array 

 

 

can anyone with some experience on drives please assist me with what i should do now?

 

will a parity check (without correction) suffice, 

 

or

 

should i trust the data and the drive (considering the partition action i performed under windows),

 

or

 

should i rather restart the drive from scratch (considering that the drive already contains about 600GB of data...)

(i am not even sure on what steps i should follow to do this... - i guess it will involve moving data to other discs, removing from array, preclearing, and then adding to array again?)

 

any assistance / recommendations would be highly appreciated

 

 

*rather safe than sorry *

 

 

Neo_x

 

 

 

 

 

 

syslog.zip

pre-clear-report.zip

[Helmonder]

Your syslog shows drive 9 to be missing... I am guessing that that is the drive you repartitioned..

 

Your array is working because of parity, the one drive is definately gone but parity makes sure you can still access all data...

 

If you stop your array you can re-assign the drive you "repaired" back to the array, it will then get written with the correct data.

 

Please note that untill you do this you are without a safetynet, loosign another drive will mean you loose data on both drives..

 

I also would advise getting a new disk, preclearing it thoroughly and make sure you have it handy for the moment your disk might fail again.. Then you can assign the new drive to replace it immediately, THEN you could do a rigoreous pre-clear of the failing drive..

[Neo_x]

Your syslog shows drive 9 to be missing... I am guessing that that is the drive you repartitioned..

 

...

 

...

 

 

I also would advise getting a new disk, preclearing it thoroughly and make sure you have it handy for the moment your disk might fail again.. Then you can assign the new drive to replace it immediately, THEN you could do a rigoreous pre-clear of the failing drive..

 

 

thx for the advise - a spare drives is definitely a better idea

 

 

it seems i somehow supplied an older syslog

 

the drive is back and working now (new syslog attached - no missing drives)

 

as a safety step i did do a reizerfsck on it, and nothing seems out of place.

 

will keep an eye on it for the time being

 

syslog2.txt