October 18, 201213 yr Hi Guys, Would be grateful for any tips or insight anyone can shed on my box. Im running 2 Hitachi deskstar 2tb , i3 8gb ram 4.7 unraid way overbuilt for the 2 drives but drive prices skyrocketed right when I built this and im not out of space yet. Today I noticed the parity drive is showing a red ball flashing After finding the syslog I am seeing many write errors. Im thinking the parity disk crashed? These drives have been spinning/sitting 11 months under pretty light use. If this disk is dead I can pop in a 3tb Hitachi and rebuild it pretty easily im guessing. This is my first unraid issue in 11 months, that's great microsoft cant do that! Thanks for any advise ! Oct 12 20:34:22 Tower kernel: handle_stripe read error: 1402999192/0, count: 1 Oct 12 20:34:22 Tower kernel: md: disk0 read error Oct 12 20:34:22 Tower kernel: handle_stripe read error: 1402999200/0, count: 1 Oct 12 20:34:22 Tower kernel: md: disk0 read error Oct 12 20:34:22 Tower kernel: handle_stripe read error: 1402999208/0, count: 1 Oct 12 20:34:22 Tower kernel: md: disk0 read error Oct 12 20:34:22 Tower kernel: handle_stripe read error: 1402999216/0, count: 1 ***************************************************************************** Alright issue solved Thanks for the good help guru's . Onto badblocking this 12month old hitachi. steps taken for resolution from red balled parity drive 2tb to 3tb upgrade 4.7unraid 1 stop the array 2 unmount the parity drive to unassigned 3 restart the array with just good data disks 4 stop the array again 5 shut down the server 6 install the 3tb drive in place of the 2tb drive , move the 2tb drive to a new port 6.5 turn server on 7 telnet into box 8 ran hdparm --yes-i-know-what-i-am-doing -N p4294963168 /dev/sda 9 stop array if needed , loaded browser and assigned 3tb drive to parity drive location , 10 started array , start failed parity drive did not come online only data disk 11 stopped array , rebooted 12 upon reboot started parity synch, 10 hour run time , running perfect half_sys_log.txt
October 18, 201213 yr Attach a SMART report. If the disk cannot output a report, it might be as simple as a a loose cable/connector/driver tray/rack connection to either the power or SATA connection to the disk. Stop the array, power down, re-seat the connectors, then attempt to again get a SMART report. If still unable, you might try a different cable or disk controller port. If still unable, odds are the drive has died. Joe L.
October 18, 201213 yr Author Here's my first smart report ever thanks a ton for the good advise! I like the part in the log where it says drive failure expected in 24 hours . After work today ill power down and re seat. This box has had a very easy life aside from one lady who tried to use it as a footstool and was promptly told to get her feet off my hard drives. Thanks Everyone have a great Thursday! smart.txt
October 18, 201213 yr No need to re-seat cables. The disk has failed, as the SMART report is describing. The failures are re-allocated sectors, meaning bad spots on the disk platters that cannot be read. There are still sectors pending re-allocation, indicating you ran out of spare sectors. RMA the disk, and keep feet off of the server. (They probably did no direct damage, but no need to make it any more dusty than needed) Joe L.
October 18, 201213 yr SMART Attributes Data Structure revision number: 16 Vendor Specific SMART Attributes with Thresholds: ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE 1 Raw_Read_Error_Rate 0x000b 085 085 016 Pre-fail Always - 8126628 2 Throughput_Performance 0x0005 135 135 054 Pre-fail Offline - 85 3 Spin_Up_Time 0x0007 136 136 024 Pre-fail Always - 421 (Average 421) 4 Start_Stop_Count 0x0012 100 100 000 Old_age Always - 114 5 Reallocated_Sector_Ct 0x0033 001 001 005 Pre-fail Always FAILING_NOW 1349 7 Seek_Error_Rate 0x000b 100 100 067 Pre-fail Always - 0 8 Seek_Time_Performance 0x0005 133 133 020 Pre-fail Offline - 27 9 Power_On_Hours 0x0012 100 100 000 Old_age Always - 4435 10 Spin_Retry_Count 0x0013 100 100 060 Pre-fail Always - 0 12 Power_Cycle_Count 0x0032 100 100 000 Old_age Always - 17 192 Power-Off_Retract_Count 0x0032 100 100 000 Old_age Always - 116 193 Load_Cycle_Count 0x0012 100 100 000 Old_age Always - 116 194 Temperature_Celsius 0x0002 222 222 000 Old_age Always - 27 (Lifetime Min/Max 17/42) 196 Reallocated_Event_Count 0x0032 001 001 000 Old_age Always - 9736 197 Current_Pending_Sector 0x0022 098 098 000 Old_age Always - 206 198 Offline_Uncorrectable 0x0008 069 069 000 Old_age Offline - 735 199 UDMA_CRC_Error_Count 0x000a 200 200 000 Old_age Always - 0 I would issue a smart -t short or smart -t long (and wait the 4 hours) just so you get a track of the errors. After that you can "try" to do a badblocks test in write mode to recover usage of the drive, but frankly, RMA'ing the drive is probably the best bet. Too many reallocated sectors. Was there a sudden power off or a bump to the machine?
October 18, 201213 yr Author In late sept, thinking 25's I had a lockup event, tried to shutdown via the webpage and the webpage froze could not access the data drives telnet'd into the box but no commands would seem to shut down the box. some commands made it look like it was starting to shutdown but after a long wait it was doing nothing. did a hard shutdown, restarted and connectivity was back, did a parity check on my small 2tb unraid I want to say it found 2 errors. Parity checks were taking an unreal amount of time, like greater than 24 hours for 2tb, seemed long and inefficient. Then yesterday I noticed a drive was red balled. I have futher goofed and ordered a 3tb drive. There seems to be a good reason I used 2tb drives in my initial build, 4.7 dont run 3tb's. These drives are from ebay so im not sure on warranty support for desk starts through hitachi but will see, I learned from that and the new 3tb has a hitachi warranty. Thanks for the advise ive got nothing to lose by running these tests it appears ill try em. Time is not of the essence but if i lose that hard drive I will be very unhappy I should probally back up irreplacable stuff but didnt want to stress the drive.
October 18, 201213 yr Reads will not stress the parity drive. It's already failing. I would take the parity off line since it's already failing and do a badblocks test in write mode. It will do a 4 pass bad blocks test. It may possible refresh the sectors. I had a 2TB drive that failed after a hard shutoff. The badblocks did seem to refresh those reallocated sectors. It's worth a shot, I would consider replacing it. If you have the 3TB drive already in house, I would set it to 2TB with an HPA and use that or upgrade to unRAID5. Then do the whole badblocks thing on the failing drive to see if it can still be useful.
October 23, 201213 yr Author Few days later i have got a hitachi 5300 3tb drive with warranty to replace the parity disk. I am going to update to the v5 rc8a unraid distro. after stopping, upgrading linux, booting up, Im reasearching but it looks like I map the new 3tb as the parity. And here's the next step I havent found info on yet, how do I rebuild the parity? Does it do this automatically? On a side note I found the profile settings, the classic rock avatars jam! thanks for the good advise , once I get the parity up and running I'll work with the "bad" drive. Im leaning towards an os failure trashing it but will see.
October 24, 201213 yr Author Well 5.0 with 1 drive assigned was a failure. Backed up to 4.7 successfully. Anyone have a tip on how to resize the 3tb drive with 2 2tb drives plugged in? Do i need to un assign and un plug the larger drivers to run a hdparm -N xxxxxxx ? The data on that one drive is critical.
October 24, 201213 yr Author im about to hdparm -N p4294963168 /dev/sdX will this know to partition the 3tb. Im trying to avoid partitioning the good 2tb, its installed and running in the same case/unraid.
Archived
This topic is now archived and is closed to further replies.