inanepunk - SSD dropping (split from Unassigned Devices thread)

[inanepunk]

Apologies if this is the wrong place for this. I'm a long time lurker, but been actively using unraid for a few months now. Looking for some assistance/pointers with an issue i am having.

 

Background

I had a spare SSD which i added to my array and was using it as a transcode location for plex. This was before i realised that it wasn't a good idea to have SSDs in the array itself. I have since removed it from the array and have been trying to mount it using UD and present it to my containers.

When i first mount the device in UD it all appears fine - i am able to format it and mount it successfully. I can successfully navigate to it in the terminal and do basic r/w test etc. all good so far.

I then configure my container to be able to use it (mounted with rw:slave option) and restart the container, i can access the disk location through the containers terminal and the application. again all good. 

 

Problem

After about an hour, the disk appears to spin down and is no longer available to the container or unraid. it is not possible to spin the drive back up, and the following is written to the disk log:

Apr 30 11:16:16 Tower unassigned.devices: Successfully mounted '/dev/sdj1' on '/mnt/disks/transcode'.
Apr 30 12:18:03 Tower kernel: sd 1:0:8:0: [sdj] tag#0 CDB: opcode=0x28 28 00 15 d5 43 38 00 00 08 00
Apr 30 12:18:04 Tower kernel: sd 1:0:8:0: [sdj] tag#0 CDB: opcode=0x28 28 00 15 d5 43 38 00 00 08 00
Apr 30 12:18:05 Tower kernel: sd 1:0:8:0: [sdj] tag#0 CDB: opcode=0x28 28 00 15 d5 43 38 00 00 08 00
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: [sdj] tag#0 UNKNOWN(0x2003) Result: hostbyte=0x00 driverbyte=0x08
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: [sdj] tag#0 Sense Key : 0x2 [current]
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: [sdj] tag#0 ASC=0x4 ASCQ=0x2
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: [sdj] tag#0 CDB: opcode=0x28 28 00 15 d5 43 38 00 00 08 00
Apr 30 12:18:33 Tower kernel: print_req_error: I/O error, dev sdj, sector 366297912
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: [sdj] killing request
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: [sdj] UNKNOWN(0x2003) Result: hostbyte=0x01 driverbyte=0x00
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: [sdj] CDB: opcode=0x35 35 00 00 00 00 00 00 00 00 00
Apr 30 12:18:33 Tower kernel: print_req_error: I/O error, dev sdj, sector 244198784
Apr 30 12:18:33 Tower kernel: XFS (sdj1): metadata I/O error in "xlog_iodone" at daddr 0xe8e2d40 len 64 error 5
Apr 30 12:18:33 Tower kernel: XFS (sdj1): xfs_do_force_shutdown(0x2) called from line 1250 of file fs/xfs/xfs_log.c. Return address = 0000000028f9350f
Apr 30 12:18:33 Tower kernel: XFS (sdj1): Log I/O Error Detected. Shutting down filesystem
Apr 30 12:18:33 Tower kernel: XFS (sdj1): Please umount the filesystem and rectify the problem(s)

 

The only way to get the disk present again in the UD interface is to restart unraid. 

I do have the default spin down time set to 1 hour, but gather this doesn't affect UD disks (I have turned this off to double check and had the same results). 
The disk passes smart checks and doesn't have any errors, its not over heating (highest recorded temp is 33c).

 

Appreciate if anyone can give me any pointers...

[trurl]

(you should always)

 

Go to Tools - Diagnostics and attach the complete diagnostics zip file to your next post.

 

From those snippets it looks like your problem is unrelated to this plugin. I may split it into its own thread after we take a look.

[inanepunk]

18 hours ago, trurl said:

(you should always)

 

Go to Tools - Diagnostics and attach the complete diagnostics zip file to your next post.

Understood - i have attached it here.

18 hours ago, trurl said:

From those snippets it looks like your problem is unrelated to this plugin. I may split it into its own thread after we take a look.

Thank you.

tower-diagnostics-20190502-0910.zip

[JorgeB]

10 minutes ago, inanepunk said:

Understood - i have attached it here.

That SSD is dropping offline:

 

Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: Power-on or device reset occurred
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: Device offlined - not ready after error recovery
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: [sdj] tag#0 UNKNOWN(0x2003) Result: hostbyte=0x00 driverbyte=0x08
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: [sdj] tag#0 Sense Key : 0x2 [current]
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: [sdj] tag#0 ASC=0x4 ASCQ=0x2
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: [sdj] tag#0 CDB: opcode=0x28 28 00 15 d5 43 38 00 00 08 00
Apr 30 12:18:33 Tower kernel: print_req_error: I/O error, dev sdj, sector 366297912
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: rejecting I/O to offline device
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: [sdj] killing request
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: [sdj] UNKNOWN(0x2003) Result: hostbyte=0x01 driverbyte=0x00
Apr 30 12:18:33 Tower kernel: sd 1:0:8:0: [sdj] CDB: opcode=0x35 35 00 00 00 00 00 00 00 00 00
Apr 30 12:18:33 Tower kernel: print_req_error: I/O error, dev sdj, sector 244198784

You can try replacing cables, though you should connect the SSDs to the onboard SATA controller (and set to AHCI) to be able to trim them.

[inanepunk]

10 hours ago, johnnie.black said:

You can try replacing cables, though you should connect the SSDs to the onboard SATA controller (and set to AHCI) to be able to trim them.

I moved it to onboard SATA and changed the power cable. So far so good. but it hasnt been in very long, ill monitor over the next hour or so. 

I have notice that it appears to power down when not in use. is that normal for SSDs in UD? 

 

[JorgeB]

11 hours ago, inanepunk said:

is that normal for SSDs in UD? 

Yes, it doesn't spin down but it enters a lower power state, though it won't make much difference on power with SSDs.