Hey all, 

 

 

Kinda freaking out a little.  Everything was fine with my unraid 6.3.5 setup.  I ordered a couple 3TB Datacenter pulls and was about to expand my array when I noticed my 6yr old 1.5TB drive had 197 pending sectors.  I went ahead and unassigned the drive and started the array so it would mark the disk missing, then I stopped the array and assigned a new 3TB disk.  Immediately after beginning the rebuild, I started receiving errors on multiple disks

 

64 Errors on the parity disk

100+ errors on the new replacement disks 

200+ errors on disk 5, my newest and seemingly good disk.

 

Its worth noting that there is a possibility that the replacement disk is in a bad physical slot.  I had a very quick failure on a drive in this slot a month or so ago and assumed it was a bad drive.  Now, I am obviously thinking bad slot.  

 

Either way, I tried stopping the array and now my web gui is unresponsive.  I am very concerned about the multiple errors on multiple disks and have decided to pause all activity until I hear back from the experts.

 

thanks

unraid-diagnostics-20171214-1003.zip

preclear_report_P8GU55TP_2017.12.13_15.51.15.txt

preclear_report_P8GV0EAP_2017.12.13_16.10.04.txt

Edited December 14, 20178 yr by wisem2540
add logs

Post your diagnostics.

I can reach the server via telnet only at the moment.  Do I just copy out the syslog?  Or is there a command I can run to capture diagnostics?  After capturing should I reboot?  I have no access to the web GUI or any shares

type diagnostics then reboot

had to send up powering off hard.  But I have the logs.

 

Attached to first post

 

Edit:  upon reboot, all drives are showing green except disk 2, the one I tried replacing.  It is now marked disabled.  However, the preclear seemed to go fine.  I will attach the preclear logs as well.  I cleared 2 drives, but only added one so far

 

Thanks

Edited December 14, 20178 yr by wisem2540

Looks like the typical SASLP issue, controller crashed taking multiple disks with it, you can try again but you should replace it with an LSI when possible, as it will likely happen again in the near or far future.

Johnnie,

 

thanks for the quick reply.  Are you referring to the supermicro controller?  Isnt this the controller that pretty much every unraid build is based off of?  Is this no longer the recommended path?  Also it is safe to say that, at this point, I have not lost any data?

 

Any elaboration you can provide would be most helpful.  Would switching to the IBM brand controller solve this issue?

 

Thanks

Edited December 14, 20178 yr by wisem2540

The SASLP worked well with unRAID before v6, since then it's been an issue for multiple users, though not all are affected, but enough for no longer being recommended, same for the newer SAS2LP which uses the same driver.

 

Current recommended controller is any LSI SAS2008/2308/3008 based controller in IT mode, SAS2008 is the older and cheapest one and enough for any HDD array, several models are available, most popular are the 9211-8i, 9201-8i and clones like Dell H310 and IBM M1015.

 

Your data should be safe, but some corruption is possible, you'd need to have checksums (or be using btrfs) to be sure all is fine.

I have easy access to a bunch of IBM 3082E-R controllers.  A quick search seems to yield some limited success with unraid.  Can you comment on those specifically?

It should work but it will be limited to 2TB and smaller disks. it's a SAS1 LSI controller based on SAS1068 chipset, for >2TB support you need a SAS2xxx or newer based controller.

You are awesome man.  Thanks a bunch.  I found some ebay M1015s for about 40 dollars.  One last question.  Because of my unclean shutdown, unraid wants to run a parity check.  I assume I should left that finish before trying to rebuild that drive?  Should I stop it and add the 1.5TB back in that I tried to upgrade?  I dont like being one drive down without dual parity

7 hours ago, wisem2540 said:

Because of my unclean shutdown, unraid wants to run a parity check.  I assume I should left that finish before trying to rebuild that drive?  Should I stop it and add the 1.5TB back in that I tried to upgrade?

It won't be a parity check with a disable drive and single parity, it will be a read check, might as well replace the failed drive before.