May 26, 201412 yr OK, someone told me the tower was down, so I checked the status, and disk1 had a red ball, and parity was blue and apparently unassigned/new. I freaked out, shut it down, and the next time I booted it up, disk 1 showed up green, but the parity is still missing. I assume that one or more drives is about to fail, but really I'm over my head and could use some help/advice. here is my syslog and smart reports. I'm unsure what else I can do without losing data. I'm running Unraid version 5.0.4 syslog.txt smartb.txt smartc.txt smartd.txt
May 26, 201412 yr The formatting in those logs is almost unreadable. I've picked through them quickly. You need to look at the results for smartb. I'm no hard disk genius but that looks knackered mate. It's riddled with errors.
May 26, 201412 yr The formatting in those logs is almost unreadable. I've picked through them quickly. You need to look at the results for smartb. I'm no hard disk genius but that looks knackered mate. It's riddled with errors. Smartb seems to indicate a cabling problem now or at some point in the past. It has no reallocated sectors or pending sectors, the major causes for disk health concerns.
May 26, 201412 yr In the Syslog, these two disks are appearing and disappearing like a David Copperfield illusion: (sdc) Hitachi_HDS722020ALA330_JK1101YBK47W9F size: 1953514552 (sdb) ST2000DL003-9VT166_5YD185YV size: 1953514552 bjp999's advice is right...check the cables...make sure they're tight (or use locking connector SATA cables if you have them around), don't 'tie' the SATA cables together (it looks neat and tidy but creates error-producing crosstalk.) And check your power supply...is it large enough for seven drives + motherboard + peripherals? And, try not to put all the drives on the same long chain of connectors...divide them up, if your power supply is wired for that. Finally, make sure all the power connectors are tight in the drives.
May 26, 201412 yr Author OK, I'll check all the cables and maybe swap the arrangement of power cords. What about my parity drive turning blue? Should I just rebuild it, or is there a more proper step?
May 27, 201412 yr Check the cables and power connectors...start up unRAID and report back with a syslog. I can't suggest what to do about parity without knowing where your disks are at. If parity WAS good, and the drives all come back online, then a parity check would be fine. BUT, if any drive is still bad, then you don't want to touch parity, yet...it may be needed to recover a failed disk.
May 27, 201412 yr Author OK, I rotated the SATA cables, and re balanced the power rails to see if that had an effect. Here is my newest syslog. Sorry I don't know how to format it better. Also, it looks like the parity drive is sdc. Additionally, drive 1 (sdb) has been giving me other problems. It was the drive that reported a red ball earlier in the week, and it had appeared to run out of space when it reported having 20+ Gigs of free space on it, and there are also some phantom files that I can't delete in disk 1, not even if I use midnight commander to delete them. Not sure if these are related symptoms or a different problem altogether, but I thought it was worth mentioning. Thanks for all the help. I really appreciate it! syslog527.txt
May 27, 201412 yr Sorry I don't know how to format it better. There was nothing wrong with the formatting of your first log as long as a viewer is used that understands linux line endings like Notepad2. When I load it with Windows Notepad then it comes out looking badly formatted because each line for Notepad is expected to end in a cr/lf pair (Carriage return/Line feed) not the linux lf (Line feed) that your files have. So nothing wrong with them. Can't tell you what your problem is by looking at the log or at least I don't want to give you bad advice so I have not looked at it just checked the formatting.
May 28, 201412 yr I don't spot the problem in the syslog, either. But if memory serves, that problem with 'more free space than the size of the drive' is sometimes a problem of extended attributes. Try the extended attributes fix mentioned in the v.5 notes. http://lime-technology.com/forum/index.php?topic=28484.0 Then, (NOT ON THE PARITY DRIVE) but on that other problem drive 1, run Reiserfsck --check to see if the file system is hosed. http://lime-technology.com/wiki/index.php?title=Check_Disk_Filesystems Both of these steps should be safe and won't break anything.
May 28, 201412 yr Sorry I don't know how to format it better. There was nothing wrong with the formatting of your first log as long as a viewer is used that understands linux line endings like Notepad2. When I load it with Windows Notepad then it comes out looking badly formatted because each line for Notepad is expected to end in a cr/lf pair (Carriage return/Line feed) not the linux lf (Line feed) that your files have. So nothing wrong with them. Can't tell you what your problem is by looking at the log or at least I don't want to give you bad advice so I have not looked at it just checked the formatting. Every day is a school day! Symptom of being a complete Windows user. I downloaded notepad++, I can read the logs now. Your syslog does not show any suspicious errors. A failed disk I was testing recently riddled my syslog with I/O errors. Now I can see the smart report for disk b more clearly bjp999 is spot on. It does look like a cabling issue, although if all cabling checks out and you try a different SATA cable and still get problems I believe I've read somewhere this can sometimes be related to miss-configured RAM. i.e. wrong timings / voltage / speed in BIOS.
May 29, 201412 yr Author Then, (NOT ON THE PARITY DRIVE) but on that other problem drive 1, run Reiserfsck --check to see if the file system is hosed. http://lime-technology.com/wiki/index.php?title=Check_Disk_Filesystems Both of these steps should be safe and won't break anything. OK, ran reiserfsck and result was "Fatal corruptions were foun, Semantic pass skipped 4 found corruptions can be fixed only when running with --rebuild-tree" running again now with --rebuild tree "Objects without names 18 Files linked to /lost+found 18 Objects having used objectids : 6 Files fixed 6 Deleted unreachable items 16 Flushing..Finished Synching..Finished reiserfsck finished at..." So far, so good? I'll run a complete S.m.a.r.t. Test, then try the extended attributes fix. Then I should be good to go, right?
Archived
This topic is now archived and is closed to further replies.