April 3, 201610 yr Not sure how much of this background is relevant to the syslog attached but I'll provide it anyway. A few weeks ago the 500GB Crucial BX100 I was running as my cache drive vanished from the array for no reason. I replaced it with a 500GB Samsung 850 EVO and rebuilt my Plex and PlexWatch docker containers from scratch. After some back and forth with Crucial support, they convinced me I could power cycle the BX100 in my Windows computer if it didn't show errors there I should be able to resume normal operation with it. And sure enough, when I did that the drive appeared to be working normally again and Stablebit Drive Scanner reported no errors. So I figured why not put it back in the unRAID server as a 2nd cache drive for redundancy? Did that, and as soon as added and formatted it as cache 2 I see all green dots by my cache shares telling me everything's protected, and everything was working fine since then until this morning when I rebooted to upgrade to 6.1.9. Now the log is showing cache drive read errors and Docker won't start. Any advice on what I need to do to recover? Thanks. jbox-syslog-20160403-0628.zip
April 3, 201610 yr Author Attached are SMART reports for both SSDs if these help too. Thanks again. CT500BX100SSD1_1453F0015320-20160403-1057.txt Samsung_SSD_850_EVO_500GB_S2RANXAH129258T-20160403-1102.txt
April 4, 201610 yr Device Model: CT500BX100SSD1 Serial Number: 1453F0015320 5 Reallocated_Sector_Ct 0x0000 100 100 000 Old_age Offline - 0 196 Reallocated_Event_Count 0x0000 100 100 000 Old_age Offline - 1 Strange, 1 reallocated event but 0 reallocated sectors, in any case the SSD should be fine. There fs corruption on your cache, try starting the array in maintenance mode and running: btrfs check --repair /dev/sdj1
April 4, 201610 yr Author Thanks Johnnie. Running that command returned only the following line that stood out to me: "cache and super generation don't match, space cache will be invalidated" Everything else just showed # of bytes but no errors being fixed "Fixed 0 roots"; "found 28810358812 bytes used err is 0", etc. After that I stopped and started the array and tried to run btrfs balance on the cache pool. I tried this yesterday too after a btrfs scrub corrected a ton of errors but every time I try balance it won't complete. I keep refreshing the screen and it looks like it's making normal progress all the way through but then when it should finish it goes right back to saying "no balance". Log from after is attached. Let me know if you have any ideas on next steps. jbox-syslog-20160404-0815.zip
April 4, 201610 yr It's normal to show "no balance" when the balance ends. Are there still errors if you run a new scrub?
April 5, 201610 yr Author No errors when I run a new scrub but I assume there's still some problem with the cache pool if I'm still not able to start Docker. Can you tell me what this in the log means? Apr 4 22:17:13 JBOX emhttp: Starting Docker... Apr 4 22:17:13 JBOX kernel: BTRFS: device fsid dd965266-df7d-4563-8042-f61125b00fe5 devid 1 transid 80799 /dev/loop0 Apr 4 22:17:13 JBOX kernel: BTRFS info (device loop0): disk space caching is enabled Apr 4 22:17:13 JBOX kernel: BTRFS: has skinny extents Apr 4 22:17:13 JBOX kernel: BTRFS (device loop0): parent transid verify failed on 351387648 wanted 80799 found 41625 Apr 4 22:17:13 JBOX kernel: BTRFS (device loop0): parent transid verify failed on 351387648 wanted 80799 found 41625 Apr 4 22:17:13 JBOX kernel: BTRFS: failed to read tree root on loop0 Apr 4 22:17:13 JBOX kernel: BTRFS: open_ctree failed Apr 4 22:17:13 JBOX logger: Not starting Docker: mount error
April 5, 201610 yr The Docker file is a BTRFS disk image. Those errors are showing it has issues. You probably have to re-build it. It's fairly simple, stop Docker, delete the image file, start Docker and then pick the my-*** templates to re-install the Dockers.
April 5, 201610 yr Author The Docker file is a BTRFS disk image. Those errors are showing it has issues. You probably have to re-build it. It's fairly simple, stop Docker, delete the image file, start Docker and then pick the my-*** templates to re-install the Dockers. Okay, tried this, Docker started, and it appeared to reinstall both my containers ("Command finished Successfully!") but then there's nothing there. In the basic view on the Dockers tab it just says "No Docker Containers Installed" and when I switch to Advanced View I see under that "orphan image" with a grayed out picture of a hard drive where the Plex and PlexWatch logos should be. Am I not going to be able to recreate these containers pointing to the appdata folder that already existed? I'd rather not have to start from scratch rebuilding my Plex server given how much work I put in to adding custom collections, tweaking the metadata, etc.
April 5, 201610 yr Are you running 6.1.3? If so update to 6.1.9 and update dynamix webgui after that Then try to reinstall.
April 5, 201610 yr Author Are you running 6.1.3? If so update to 6.1.9 and update dynamix webgui after that Then try to reinstall. That did the trick. Thanks!
Archived
This topic is now archived and is closed to further replies.