PatInTheHat Posted January 4, 2012 Posted January 4, 2012 I upgraded my parity drive from a 1TB to a 2TB and everything went smoothly. Several parity checks and smart reports on all drives looked good. I also had a data disk that had been slowly failing, see thread: http://lime-technology.com/forum/index.php?topic=16059.0 which I finally replaced with a 2TB drive as well. Everything appeared to go smoothly but last night I ran my first parity check with the new drive and found 2 parity errors almost immediately. I am running 4.7 and as I understand there was a bug that caused parity errors when upgrading to a larger capacity disk and I believe this is what I have encountered. However I have no idea how to proceed or what version of unraid I should upgrade to. (I cannot find the 4.7.1 version that fixed this error and I don't want to upgrade to a possible buggy beta version) Appreciate any help, thanks.
evans036 Posted January 4, 2012 Posted January 4, 2012 if i remember correctly, the 4.7 bug may corrupt data if data is written to the array whilst a data drive is being rebuilt from the parity drive. were you writing data to the array whilst the new data drive was being rebuilt? if not, you need to figure out if it is the parity drive at fault or is it the data drive. you might want to try a reiserfsck on the data drive. perhaps smart reports could indicate drive issues. look in the syslog for drive related IO erros too if you still have your old data drive, maybe do a hash verification using md5deep using the hash file from one drive to verify the hash on the other. hope some of this helps, steve
PatInTheHat Posted January 4, 2012 Author Posted January 4, 2012 Yikes just had a look at my syslog and it's jammed full of errors. Attached below Sadly i've already run a preclear on the old disk. I have no idea how to proceed from here syslog-2012-01-05.zip
PatInTheHat Posted January 5, 2012 Author Posted January 5, 2012 *facepalm* sda isn't in the array any more. It's the drive I replaced above. It's currently being housed in the system while I run preclears on it but in no way is assigned to my array. Why on earth would a parity check cause sys errors for a drive not in the array? As a side note I am currently running another parity check (without error correction) and it did not error in the spot my previous parity check ran. (no errors so far) I originally ran the default parity check, would this have corrected those errors automatically? If not where the heck did those errors go?
dgaschk Posted January 5, 2012 Posted January 5, 2012 sda is filling your log. Remove it and then collect a new log.
evans036 Posted January 5, 2012 Posted January 5, 2012 As a side note I am currently running another parity check (without error correction) and it did not error in the spot my previous parity check ran. (no errors so far) I originally ran the default parity check, would this have corrected those errors automatically? If not where the heck did those errors go? perhaps an intermittent problem caused by bad cable?
PatInTheHat Posted January 5, 2012 Author Posted January 5, 2012 As a side note I am currently running another parity check (without error correction) and it did not error in the spot my previous parity check ran. (no errors so far) I originally ran the default parity check, would this have corrected those errors automatically? If not where the heck did those errors go? perhaps an intermittent problem caused by bad cable? I recently re-seated all my cables but I will have another look. Attached is my latest smart report without the sda drive. Everything looks fine to me, anyone mind confirming? I have run 2 parity checks and everything is coming up fine. Smart reports also reporting perfectly. Should I just ignore those 2 parity errors and continue as if they never happened? syslog-2012-01-05.zip
evans036 Posted January 5, 2012 Posted January 5, 2012 i would probably do a reiserfsck --check /dev/md1 but replace md1 with the data disk that threw the parity errors this would just give another level of comfort that the data drive is ok. good luck, steve
PatInTheHat Posted January 6, 2012 Author Posted January 6, 2012 i would probably do a reiserfsck --check /dev/md1 but replace md1 with the data disk that threw the parity errors this would just give another level of comfort that the data drive is ok. good luck, steve Very newbie question but how do I know which drive produced the errors? They didn't show up under any particular drive from the scan, it just said 2 parity sync errors in the final results.
lionelhutz Posted January 6, 2012 Posted January 6, 2012 You generally can't know which drive had the issues if there isn't some obvious errors being thrown by one of the drives. All you know is that the parity calculation using the data drives didn't match what was stored on the parity disk. If the parity errors were at the very beginning of the check then I wouldn't worry about it. But then there's not much you can do anyways at this point. I expect it was the 4.7 rebuild bug. Peter
evans036 Posted January 6, 2012 Posted January 6, 2012 i would probably do a reiserfsck --check /dev/md1 but replace md1 with the data disk that threw the parity errors this would just give another level of comfort that the data drive is ok. good luck, steve Very newbie question but how do I know which drive produced the errors? They didn't show up under any particular drive from the scan, it just said 2 parity sync errors in the final results. i have never had a parity sync error (yet) - but i just assumed that the syslog would have identified the drive that threw the parity error. no? but it does sound like your system is now solid. i would at this point perform regular parity checks and otherwise assume all is well. good luck, steve
Recommended Posts
Archived
This topic is now archived and is closed to further replies.