September 30, 201213 yr I built the system,as recommended, started adding drives everything was up and running, 11 drives+ parity upgraded Unraid version upon every release. In the version where root user was removed from samba shareto only work with console and web ui account, had a lot of issues with file permissions the problems from there continued i ran a ReiserFS check on all drives, to see if the permission problems i had are caused because of hardware, reiserfs check all pass, but md2 (Disk2) ,failed tried to fix it (fix) - did not work tried to rebuild it (rebuild) - did not work failed, and i had to format the drive ( lost 3tb of data ! ) [2012.07.20 screenshot] Since then - failed parity check all the time the system freezes up, telnet is responsive but not functional samba share is showing but freezes on access web ui is down (failed the parity check about 20 times in a period of approximately one month) requires force reboot or shutdown to start working (sometimes 2 to identify all of the drives). (all drives pasS s.m.a.r.t test - upon post, every time,even now) When the array is up and running (normal operation) access to samba is working (read+write) web ui working spin up, spin down working (no power issues) [2012.09.22 screenshot] my steps to fix the problem: switched the motherboard - works with new MB but failed parity check (from biostar to asus) switched the ram - works with new ram but failed parity check (from Samsung 2gb to ocz 1gb) switched the CPU - works with new cpu but failed parity check (dual core to core2Duo) ReiserFS - pass on all drives - no Corruption found switched version from 5.0-rc5a to 5.0-rc8a - works, failed parity check did nor freeze but disk1 was missing (after 65% parity check) after reboot - disk1 is disabled, unraid works in degraded mode [2012.09.22 screenshot] Tried to transfer as many files as i could to other drives in the array at a certain point could not, rebooted to try again drive appeared as unfomatted [2012.09.26 screenshot] All data that was there, and that i saw in degraded mode, was not there, i did not format the drive tried to stop and start the array, reboot, change sata connection, did not work then stopped the array, chose no device in drop down menu of disk 1 then started the array, then stopped, chose the drive again the color changed to blue, and i had an option to rebuild, i chose it [2012.09.27 screenshot] ran rebuild data, system freeze after 40% - force reboot changed power cable arrangement for drives (different cable) ran rebuild data, fail, output on the monitor was IO error reiser fs error device md11 [2012.09.28 screenshot] changed MB to biostar G31-M7 TE changed psu to Antec NEO ECO 400C changed cable arrangement and power distribution changed DOK to sandisk cruzer 2gb changed ram system did nor recognize nic reverted back to - unRAID Server 5.0-rc5 AiO rebuild data option is missing (probably data is gone!) disk1 is Unformatted [2012.09.29 screenshot] running parity check - FAILED , no samba, telnet is responsive but not functional [2012.09.30 syslog] hardware: Antec Neo ECO 520C Asus P5W64 WS Professional interl core2 duo 6600 CORSAIR TWIN2X2048-6400 (one stick) sandisk cruzer 2gb Supermicro AOC-SASLP-MV8 3ware CBL-SFF8087OCF-05M 0.5M (cables) drives: see screenshot 2012.09.30_syslog.txt
September 30, 201213 yr The Antec NEO ECO 400C is not sufficient for 12 drives. The 520C should work but it's close the the limit. Which PSU is installed now? Attach all SMART reports.
October 1, 201213 yr Several disks have pending sectors. The sectors are unreadable and will prevent the successful rebuild of any disk. Usually pending sectors are corrected by rebuilding the problem disk. DGA:2012 dga$ grep "Current_Pending_Sector" * smart1-sda.txt:197 Current_Pending_Sector 0x0032 200 198 000 Old_age Always - 4 smart10-sdj.txt:197 Current_Pending_Sector 0x0012 100 100 000 Old_age Always - 0 smart11-sdi.txt:197 Current_Pending_Sector 0x0032 252 252 000 Old_age Always - 0 smart12-sdh.txt:197 Current_Pending_Sector 0x0012 100 100 000 Old_age Always - 8 smart2-sdk.txt:197 Current_Pending_Sector 0x0032 200 200 000 Old_age Always - 0 smart3-sdm.txt:197 Current_Pending_Sector 0x0032 200 200 000 Old_age Always - 1 smart4-sdc.txt:197 Current_Pending_Sector 0x0012 100 100 000 Old_age Always - 0 smart5-sdl.txt:197 Current_Pending_Sector 0x0012 100 100 000 Old_age Always - 0 smart6-sdf.txt:197 Current_Pending_Sector 0x0012 200 200 000 Old_age Always - 0 smart7-sdd.txt:197 Current_Pending_Sector 0x0032 200 195 000 Old_age Always - 14 smart8-sdg.txt:197 Current_Pending_Sector 0x0032 100 100 000 Old_age Always - 0 smart9-sdb.txt:197 Current_Pending_Sector 0x0012 100 100 000 Old_age Always - 0 2 disks have reallocated sectors. This is not an issues as long as these values remain stable. DGA:2012 dga$ grep "Reallocated_Sector_Ct" * smart1-sda.txt: 5 Reallocated_Sector_Ct 0x0033 200 200 140 Pre-fail Always - 0 smart10-sdj.txt: 5 Reallocated_Sector_Ct 0x0033 100 100 036 Pre-fail Always - 0 smart11-sdi.txt: 5 Reallocated_Sector_Ct 0x0033 252 252 010 Pre-fail Always - 0 smart12-sdh.txt: 5 Reallocated_Sector_Ct 0x0033 100 100 010 Pre-fail Always - 0 smart2-sdk.txt: 5 Reallocated_Sector_Ct 0x0033 200 200 140 Pre-fail Always - 0 smart3-sdm.txt: 5 Reallocated_Sector_Ct 0x0033 200 200 140 Pre-fail Always - 0 smart4-sdc.txt: 5 Reallocated_Sector_Ct 0x0033 100 100 036 Pre-fail Always - 14 smart5-sdl.txt: 5 Reallocated_Sector_Ct 0x0033 100 100 036 Pre-fail Always - 14 smart6-sdf.txt: 5 Reallocated_Sector_Ct 0x0033 200 200 140 Pre-fail Always - 0 smart7-sdd.txt: 5 Reallocated_Sector_Ct 0x0033 200 200 140 Pre-fail Always - 0 smart8-sdg.txt: 5 Reallocated_Sector_Ct 0x0033 100 100 036 Pre-fail Always - 0 smart9-sdb.txt: 5 Reallocated_Sector_Ct 0x0033 100 100 036 Pre-fail Always - 0 Several disks show power off retracts. Disk one in particular is having power issues every 2.5 hours. issues DGA:2012 dga$ grep Power-Off_Retract_Count * smart1-sda.txt:192 Power-Off_Retract_Count 0x0032 198 198 000 Old_age Always - 1568 smart11-sdi.txt:192 Power-Off_Retract_Count 0x0022 252 252 000 Old_age Always - 0 smart2-sdk.txt:192 Power-Off_Retract_Count 0x0032 200 200 000 Old_age Always - 46 smart3-sdm.txt:192 Power-Off_Retract_Count 0x0032 200 200 000 Old_age Always - 56 smart6-sdf.txt:192 Power-Off_Retract_Count 0x0032 199 199 000 Old_age Always - 752 smart7-sdd.txt:192 Power-Off_Retract_Count 0x0032 200 200 000 Old_age Always - 423 DGA:2012 dga$ These problems are probably due to an insufficient power supply. Try a PSU with at least 50 Amps on the 12V rail.
October 1, 201213 yr Author First of all, thank you for your assistance, The reports i outputted, was done using the antec neo-eco 520c. Let me understand What is the meaning of your analysis: The antec neo eco is recommended by unraid community and has 40A on the 12V rail, is that not enough ? the server was working fine for a long while with 12 disks on, are you saying that if i will replace the PSU , all will be well?
October 1, 201213 yr First of all, thank you for your assistance, The reports i outputted, was done using the antec neo-eco 520c. Let me understand What is the meaning of your analysis: The antec neo eco is recommended by unraid community and has 40A on the 12V rail, is that not enough ? the server was working fine for a long while with 12 disks on, are you saying that if i will replace the PSU , all will be well? Thats's what I'd try. However fixing all of the pending sectors may be difficult. Copy the data of of one drive at a time and then preclear and reassign the drive. This needs to be repeated once for each drive with pending sectors. The PSU thread says 3 Amps per drive plus 5 amps for the MB. This totals 41 amps for 12 drives.
October 1, 201213 yr Author I am officially desperate, i knew this was a complex system, but did not imagine the trouble and time that it consumes, I did contacted the "official" support about a week ago, because i have many question, but did not received any response. maybe you can answer some of them: 1) could i ,by any chance , save the data on the "Unformatted" drive ? 2) are the solutions you gave me are final and absolute, or after attended, could be eventually caused by other issues? 3) what is the procedure to preclear and reassign those drives, that are functional as part of the array? 4) what does "pending sectors" mean? could "Power-Off_Retract_Count" be caused by an Inline SATA Connector? 5) if i will replace two or three of the regular drives with green drives will it solve the power problem and will it rebuild correctly since the system is unstable? again,thanks.
October 4, 201213 yr Author Moved data from 2 500gb hard drives and disconnected them from the array Then used the "new config" utility, assigned all drives again correctly (by screenshot) [2012.10.04(1)-screenshot] [2012.10.04(2)-screenshot] Started the array, ran parity sync (automatically) Sync fail, UI down, samba freeze, try to output syslog, force restart [2012.10.04-syslog] This was suppose to work,The PSU power should be sufficient, help anyone? 2012.10.04-syslog.txt
October 4, 201213 yr Every disk with a pending sector will give read errors. Copy data off of those drives and then use pre-clear to fix the pending sectors.
October 4, 201213 yr Drives get bad sectors all the time. It's impossible to manufacture a hard drive that contains zero bad sectors. We don't typically see those because they're mapped out at the factory. Sometimes, sectors go bad after we purchase the drive. The drive is supposed to move the bad sector to a pool of reserved sectors but sometimes it's not instantaneous. So yes, it could happen again.
October 4, 201213 yr Looking at your h/w config, it appears you have 15 total SATA ports: - 7 on the motherboard - 8 on the controller card In the latest system log it appears the failure is on a drive attached to the controller card. To figure this out you need to isolate hardware. Here's what I would do: 1. Create a 7-drive array with Parity and 6 data drives, all connected to ports on the motherboard. Execute a parity-sync, then parity check. 2. If this passes, connect all the drives to the controller card. Using unRaid version 5, there should be no config changes necessary - when you power up it should just recognize all the drives have been attached to new ports. Now run a parity check. 3. If you get through both steps above, well time to add more drives, but must be done in a controlled manner. For example, move the above 7 back drives to your "desired" final SATA ports, and run a parity check to convince yourself it's still working. Then unassign parity, add a single data drive, re-assign parity, and run a parity sync followed by parity check. If you find a configuration that fails, then move the offending drive to a different (unused) SATA port and repeat test. I'm sorry if this seems tedious, but it appears there is a marginal piece of h/w somewhere and there is no easier way to isolate. If you get to a point where a set of drives fails no matter what combination of ports are used, probably something wrong with power - are you using any "Y" power connectors? These have been known to cause issues in the past. Your h/w spec does not mention the physical construction of your server - are you using drive cages, or cabling directly to the drives? Also use -rc8a.
October 5, 201213 yr Author Currectly i am using: Antec Neo ECO 520C Biostar G31-M7 TE intel dual core Kingston 2GB ram sandisk cruzer 2gb Supermicro AOC-SASLP-MV8 3ware CBL-SFF8087O Edimax gigabit NIC (which "rc8a" did not recognized) -Drives are in cages, connected via inline sata cable (no splitters). -Did tried lots of power variations, with same results. -Did tried Cabling variations, with same results. just a reminder, all started when a working 12 disk unraid system started to fail parity checks, UI failed, Telnet, Samba and Console nonfunctional, Then came more trouble... It appears that "dgaschk" and "limetech" have a different approach. Since i already started a preclear cycle for a drive, I will continue this for all drives, then will try to build parity then sync it. If failed, i will try limetech troubleshooting scenario. Any way you look at it, this is a S*** STORM. I will post results when i have them.
October 5, 201213 yr Author Well aware of that , will preclear and move data, preclear and move data... Preclear the parity drive as well, just to rule out hard drive integrity. Again, will take some time.
October 5, 201213 yr You don't need to copy the data back onto the drives to check parity, nor should you have to preclear them multiple times.
October 5, 201213 yr One other thing to check, relating to the age of your motherboard: capacitors. I looked up the manual for that board and got a printed date of 2008, so it seems fairly new and not as likely to have this problem. That being said, all manner of wonky and tedious problems occur when the electrolytic capacitors feeding your CPU fail. Eventually you'll get boot errors when yesterday it worked just fine... but enough of that. Just for your own peace of mind, take a look at the brown, relatively skinny and tall capacitor cans around the outside edges of your motherboard. There are some solid aluminum ones near your CPU that should pose no problems, but nonetheless, verify that all capacitors have perfectly cylindrical shapes and flat tops with no bulging or spewing out of their guts to be seen. If you notice any bulging on the can tops of those tall brown plastic-wrapped capacitors, you can be assured that you're in need of a new motherboard, or at least someone skilled who can replace the faulty capacitors. It's likely not the problem, but always worth confirming with a quick visual inspection.
October 5, 201213 yr Only the 4 drives with pending sectors need to be cleared. The pending sectors should go to zero. If pending sectors remain after one pre-clear then another is needed or RMA.
October 5, 201213 yr Author Again, different approaches to an unknown problem. PC Tech 101, rule out. I will run one cycle of preclear for each drive in the array, starting with 2X3tb drives. They will be assigned, after preclear, as parity+disk1. Then migrate data each time, to "tested" drives and prepare another drive (empty one) in the array for preclear. I must rule out disk integrity, If problem is solved, all is well!, if not, on to the next phases. I tend to agree with "dgaschk" and hope the "preclear roundabout" will solve the issue, regardless i plan to upgrade MB+CPU+RAM.
October 30, 201213 yr Well aware of that , will preclear and move data, preclear and move data... Preclear the parity drive as well, just to rule out hard drive integrity. Again, will take some time. How are you getting on ?
October 30, 201213 yr Author Well , it was the drives R/W problems, currently out of the 12 drives system i had , only 4 connected, They have been tested, precleared, synced and checked , more drives are pending data migration and testing. will update when done.
Archived
This topic is now archived and is closed to further replies.