June 23, 201214 yr Author ok thanks ill start the preclear now and post results OK so i did some reading and here is what i gathered and maybe i misinterpreted everything lol but it looks like that if your drive is jumpered you need to run a different preclear code...unless im reading this wrong if you user -a, you would have specified sector 63. if -A, sector 64. If jumpered you want sector 63. you can use preclear_disk.sh -t /dev/sdX to see if (and how) they are set. If they are precleared, you can use preclear_disk.sh -C 64 /dev/sdX or preclear_disk.sh -C 63 /dev/sdX to convert from one to the other.. (as long as you did not yet assign them to the array just yet) since i am wiping the drive and i have the newest unRAID OS i will remove the jumper *keep in mind anyone who reads this removing the jumper and re adding to your system can/will cause a loss of data* this is something i am prepared for. I just dont want to have to deal with these jumpers if at all possible.
June 24, 201214 yr With a jumper you will use the -a option. This is the only case where -a is beneficial.
June 25, 201214 yr Author With a jumper you will use the -a option. This is the only case where -a is beneficial. thanks i took off the jumper and just did the normal preclear scan...im running more smart test right now but i dont know if you guys want to see them...here is the preclear results.. preclear_results.txt
June 25, 201214 yr It does not appear to me that it cleared that one sector, so unless you get a perfect result from a long SMART test, I would rule out this drive for UnRAID use.
June 25, 201214 yr Author ok thats what i kind of figured would happen...ill post long & short smart test results when its done..thanks again!
June 26, 201214 yr Author i ran preclear on it and those are the smart results that i ran directly after the pre clear...sounds like needs to be replaced i guess
June 26, 201214 yr Author before i start a new thread. do you think unmenu can be affected by the bad disk?? i can no longer connect to server via unmenu...
June 27, 201214 yr before i start a new thread. do you think unmenu can be affected by the bad disk?? i can no longer connect to server via unmenu... Not directly effected. Unmenu will wait forever for a command to complete that it called. Telnet in and restart it. /boot/unmenu/uu I think.
June 27, 201214 yr Author before i start a new thread. do you think unmenu can be affected by the bad disk?? i can no longer connect to server via unmenu... Not directly effected. Unmenu will wait forever for a command to complete that it called. Telnet in and restart it. /boot/unmenu/uu I think. looks like that worked...took several minutes after running the command though...but now im in it shows Last parity check 15518 days ago with no sync errors. 8 not sure what thats about...just yesterday it said 9 days ago... im just going to remove disk 3 RMA it and add a new drive...on the new drive as long as it says the preclear scan passed thats all i need to do to prep it for adding it to array correct?
June 27, 201214 yr ok here are short and long smart test for disk 3 Good news! The last SMART report shows that the Preclear (and/or the last SMART long test) did fix that last sector! And cleared the one Uncorrectable sector that had been counted! It did not clear the one Current Pending Sector, and everything I know about SMART and bad sectors says it should have. (Obviously I have more to learn, or it's a small bug in the SMART firmware.) But this SMART report says the long test "Completed without error", the first full test to do so, without a read failure. You might try adding it back to UnRAID and checking the syslog to see if all those errors appear any more, and that the drive maintains full speed (UDMA/133). Note for future reference - if you run the long SMART test, there is no need to run the short test, because what it tests is included in the long test.
June 27, 201214 yr Author thank you @RobJ i got frustrated the drive is under warranty so i already set up an RMA oh well lol ill keep that in mind with the short and long test ha ha sorry...thanks for your help sorry i jumped the gun on the RMA for this drive (= should have waited for your response
June 27, 201214 yr ok here are short and long smart test for disk 3 It did not clear the one Current Pending Sector, and everything I know about SMART and bad sectors says it should have. (Obviously I have more to learn, or it's a small bug in the SMART firmware.) No smart test will fix a sector pending re-allocation, as that would require writing the new sector. Since the old sector cannot be read, itdoes not know what to write. The sector pending re-allocation will stay that way until that sector is written to by the operating system. At that point most disks will first try to re-write it in place, and not re-allocate it, and if that fails, re-allocate it, since it then knows what should be written.
June 27, 201214 yr Author No smart test will fix a sector pending re-allocation, as that would require writing the new sector. Since the old sector cannot be read, itdoes not know what to write. The sector pending re-allocation will stay that way until that sector is written to by the operating system. At that point most disks will first try to re-write it in place, and not re-allocate it, and if that fails, re-allocate it, since it then knows what should be written. so does that mean the drive is no good? at least for unraid?
June 27, 201214 yr No smart test will fix a sector pending re-allocation, as that would require writing the new sector. Since the old sector cannot be read, itdoes not know what to write. The sector pending re-allocation will stay that way until that sector is written to by the operating system. At that point most disks will first try to re-write it in place, and not re-allocate it, and if that fails, re-allocate it, since it then knows what should be written. so does that mean the drive is no good? at least for unraid? no, it indicates you have not written to the sector since it was marked for re-allocation. Most modern disks have several thousand spare sectors to be used for re-allocation. To have only 1 defective is not really an issue, as long as others do not continue to be marked as un-readable.
June 27, 201214 yr i see...well thanks again for everything I'd run it through several more full preclear disk cycles, or badblocks cycles and see if the one sector is the only one that shows itself. If others exist, RMA it. If not, you can RMA it, but run the risk of the replacement being no better.
June 27, 201214 yr Author not finding an easy way to cancel my RMA so ima just bite the bullet and send it in
June 27, 201214 yr not finding an easy way to cancel my RMA so ima just bite the bullet and send it in Most RMA's self cancel if you don't send it in, unless you opted for the advanced replacement option. They will bill you for the drive they send if you don't return the old drive. For a period of time after the flood, it was cheaper to buy the advanced RMA drive than buying a refurb on the open market. Look in your RMA paperwork to see if the RMA is only valid for a certain number of days. If so, and you haven't given them a CC number, they just cancel the RMA if you haven't sent it in that period of time.
June 28, 201214 yr Author i did the advanced RMA oh well lol ill preclear the new one and hope its all good!
June 28, 201214 yr ok here are short and long smart test for disk 3 It did not clear the one Current Pending Sector, and everything I know about SMART and bad sectors says it should have. (Obviously I have more to learn, or it's a small bug in the SMART firmware.) No smart test will fix a sector pending re-allocation, as that would require writing the new sector. Since the old sector cannot be read, itdoes not know what to write. The sector pending re-allocation will stay that way until that sector is written to by the operating system. At that point most disks will first try to re-write it in place, and not re-allocate it, and if that fails, re-allocate it, since it then knows what should be written. Joe, I don't know if you realized that he had just finished a Preclear on the drive (and posted the successful Preclear report), before running a short and long test to check the drive one more time. That long test showed that the Preclear must have fixed the one bad sector, because for the first time it completed the full test of the media surface, and without any read failures at all. The drive now looks fine, *except* that the Current Pending sector count never changed from a value of one. It did drop the Offline_Uncorrectable count from one to zero, something I have never seen before, since I thought that was an event count, which should only increase. It makes me wonder if the silly drive cleared the wrong count.
June 28, 201214 yr It makes me wonder if the silly drive cleared the wrong count. Situations like this are what makes me scared to always trust that the drive will do the right thing when it's asked to participate in a parity correction situation. I would rather all drives go into data lockdown read only mode until I can figure out what went wrong, or at least attempt to figure it out.
July 7, 201213 yr Author ok so here is what ive been doing....I got a replacement disk for disk3 whether it was needed or not oh well...i ran pre clear 3 times and came back successful each time. also during that time i upgraded disk 7 from 1TB to 2.5TB hdd also did 3 pre clear scans on the disk and came back fine. here is a syslog i want to confirm everything is fine now...thanks! syslog-2012-07-07.zip
July 9, 201213 yr ok so here is what ive been doing....I got a replacement disk for disk3 whether it was needed or not oh well...i ran pre clear 3 times and came back successful each time. also during that time i upgraded disk 7 from 1TB to 2.5TB hdd also did 3 pre clear scans on the disk and came back fine. here is a syslog i want to confirm everything is fine now...thanks! This syslog looks pretty good, but there are a few glitches. The rebuild of Disk 7 went fine, but about 2 and a half minutes into it, there was an longer than normal write time on the new Disk 7, and about an hour and a half later, another long write occurred. No more writing issues occurred, so the write retries must have occurred successfully. About a half hour later, Disk 1 found 2 bad sectors (Media error, with UNC flag). Disk 3 must have previously been rebuilt successfully, there are no issues present. Check SMART reports for Disk 1 and Disk 7. Compare these reports with the previous ones for each. Then run a full parity check, to be sure all is well with Disk 1 and Disk 7.
Archived
This topic is now archived and is closed to further replies.