October 3, 20178 yr Version: 6.5.3 I am experiencing random crashes with the server where it takes itself offline. There does not seem to be any logical pattern to it. As of right now I cannot get through a parity check without the array restarting. I have ran memtest for about 6 hours and there were zero errors found. I ran FCP diagonstic mode since the last crash and it made it about 12 hours again until it crashed. I have uploaded the first diagnostic and the last before it crashed. I have also included the FCP tail log. Appreciate any support. unraid-diagnostics-20171002-0923.zip FCPsyslog_tail.txt unraid-diagnostics-20171001-1949.zip
October 3, 20178 yr Not sure I can help solve your problem or not but here is what I found: The following disks are connecting at PATA speeds (so really SLOOOOOW) unless they ARE PATA disks (don't think WD30EFRX are). You need to change a BIOS setting to make them connect at SATA/AHCI: WDC WD30EFRX-68EUZN0 WD-WMC4N1500053, WDC WD30EFRX-68EUZN0, WD-WMC4N0K1PUUR. You also have ReiserFS disks on your system. unRAID v6 doesn't like ReiserFS disks when you write to them and can cause weird issues (not sure spontaneous crashes are one or not). If you set all shares on the ReiserFS disks to be read only you may be able to migrate them to XFS which I also see some of your disks are in XFS format. So if you have the same share on XFS and ReiserFS disks then you will have to make them all Read/Only until you have migrated all data off of ReiserFS. An alternative would be to just remove all the ReiserFS disks from the system and mount them one at a time and copy the data off them to your XFS only array with the Unassigned Devices plugin. That's what I found and hopefully that helps or someone else maybe gets you better advice.
October 4, 20178 yr Author Thank you for the input. No hardware I have should be PATA.... I will look into my BIOS and investigate if I can change from PATA to SATA. (I'm assuming I can stumble my way through it) Funny thing about the ResiserFS... I recently tried to convert from Reserfs to XFS based upon a Sticky guide. I completely overlooked the part about migrating your data first... WHOOPS. No clue what I was thinking. I lost 12TB worth of data before I caught the mistake. Thankfully Johnnie Black was able to guide me through a recovery process where from what I can tell I was able to recover 100%. Woot! I will make them ready only and then work on migrating the data. AND THEN CONVERT.
October 4, 20178 yr That sounds like the correct procedure. I love the ReiserFS recovery tools and wish the XFS and BTRFS recovery tools were as good. I have experience with a ReiserFS recovery myself after trying to building parity on a full data/cache drive for ~5 minutes. I got back all but 200GB of a 2TB ReiserFS drive when I ran a recovery.
October 4, 20178 yr Author Do you see any data concerns with switching my BIOS from IDE to AHCI? This should resolve my PATA issue but I don't want to compromise my data on the discs in the process. I am inclined to think no, since its really just about emulation from what I can tell.
October 4, 20178 yr It shouldn't cause a problem as far as I know. I don't believe it changes how sectors are accessed by the OS just the method of transport to and from the MB and drive.
October 9, 20178 yr Author Update -- so I successfully migrated all my data around and converted everything to XFS. I kept my fingers crossed that would do the trick. I attempted to rebuild the parity and twice now it has crashed before even getting to 3%. The second time around I had the FCP logger going. Logs attached. Where do I go from here? I'm not sure what to do as I can't even rebuild my parity now.. Hopefully the logs yield more clues to someone as I have absolutely no idea. Replace the memory? Replace the PSU? I hate to start buying and replacing parts with wild shots in the dark but this has been quite troublesome for some time now. FCPsyslog_tail.txt unraid-diagnostics-20171008-2327.zip unraid-diagnostics-20171008-2357.zip
October 9, 20178 yr Community Expert There's nothing on the logs about the crash, but you do have SATA ports 5 and 6 in IDE mode, enter the bios and change them to AHCI/SATA, not the first time I've seen this causing various issues on similar AMD boards.
October 15, 20178 yr Author What should I be looking for in the log to confirm if I have successfully turned off IDE mode?
October 15, 20178 yr Author I noticed this error pop up on the monitor while the system was running. I'm wondering if this has something to do with the server crashing?
October 15, 20178 yr Community Expert 7 hours ago, SuperNotSupper said: What should I be looking for in the log to confirm if I have successfully turned off IDE mode? The onboard controller should show 6 AHCI ports instead of four Oct 8 23:25:09 unRAID kernel: ahci 0000:00:11.0: version 3.0 Oct 8 23:25:09 unRAID kernel: ahci 0000:00:11.0: AHCI 0001.0100 32 slots 4 ports 3 Gbps 0xf impl SATA mode Oct 8 23:25:09 unRAID kernel: ahci 0000:00:11.0: flags: 64bit ncq sntf ilck pm led clo pmp pio slum part ccc These should change from pata to AHCI Oct 8 23:25:09 unRAID kernel: scsi host7: pata_atiixp Oct 8 23:25:09 unRAID kernel: scsi host8: pata_atiixp In doubt post new diags. 5 hours ago, SuperNotSupper said: I noticed this error pop up on the monitor while the system was running. I'm wondering if this has something to do with the server crashing? Most likely a consequence of the server crashing, you need to check filesystem on disk6: https://wiki.lime-technology.com/Check_Disk_Filesystems#Drives_formatted_with_XFS
October 15, 20178 yr Author It seems like I can't win with this lately. Now I have lost Disk 7 for whatever reason. Earlier today the array went crazy and did not respond to either the web GUI or telnet (monitor was just scrolling through text). So I had to do a physical reset. I powered the system back on and disk 7 was missing. I powered off, and checked all the connections and powered back on. Disk 7 was found but it is now telling me that the contents are emulated and the device is disabled. This seems extra precarious because the parity was not built because of it crashing every time before it would finish. Am I SOL with the data that was on this drive? This also begs the question if the PCI-express controller is the culprit as the last two drives are the ones being controlled by this causing the Pata issue. unraid-diagnostics-20171015-1629.zip unraid-syslog-20171015-1631.zip Edited October 15, 20178 yr by SuperNotSupper
October 16, 20178 yr Community Expert Assuming nothing was written to disk7 after being disabled, you can do a new config and re-sync parity, before that replace disk7 cables to rule them out.
October 20, 20178 yr Author I'm really baffled... I double checked all the connections. Ran a new config and re-synced the parity expecting the worst. To my surprise it successfully completed and the server hasn't crashed once in 4-5 days. I'm going to keep my fingers crossed for now that it has self corrected. The only thing I could maybe speculate is that disc 7 connection was loose. They are extremely sloppy fitting on the PCI-express controller card.
Archived
This topic is now archived and is closed to further replies.