Hello all

I seem to be having a issue with the cache drive (PCIE NVME drive WD Black 256) it seems to be unmounting or goes missing something don't know. its happened 3 times so far i can reboot the system and it comes back i was able to grab the log on the 2nd time. i forget to grab it before reboot. but i went ahead and bought a Samsung EVO 960 to replace the drive  found a good deal on amazon.but just want some one that knows how to read the log to look at it and tell me something.maybe the drive is going bad its only around 3 mouth old so maybe not don't know good thing it has a 5 year warranty. almost for got to tell you the system its a HP Elite 8300 core I7 3770 16 gig ram 3x 3TB WD red's and 1x 256 GB WD Black PCIE NVME with a APC back up Pro 1500.

tower-syslog-20171103-0451.zip

Edited November 4, 20178 yr by Gary489

Next time grab the complete diagnostics: Tools -> Diagnostics.

57 minutes ago, Gary489 said:

This help? i grab the current and one i had

 

tower-diagnostics-20171104-1917.zip

tower-diagnostics-20171103-0453.zip

It's definitely a hardware problem:

 

Nov  3 02:01:41 Tower kernel: nvme nvme0: I/O 140 QID 8 timeout, aborting
Nov  3 02:01:41 Tower kernel: nvme nvme0: I/O 146 QID 8 timeout, aborting
Nov  3 02:01:41 Tower kernel: nvme nvme0: I/O 150 QID 8 timeout, aborting
Nov  3 02:01:41 Tower kernel: nvme nvme0: I/O 154 QID 8 timeout, aborting
Nov  3 02:01:41 Tower kernel: nvme nvme0: I/O 155 QID 8 timeout, aborting
Nov  3 02:01:41 Tower kernel: nvme nvme0: Abort status: 0x0
Nov  3 02:01:41 Tower kernel: nvme nvme0: Abort status: 0x0
Nov  3 02:01:41 Tower kernel: nvme nvme0: Abort status: 0x0
Nov  3 02:01:41 Tower kernel: nvme nvme0: Abort status: 0x0
Nov  3 02:01:41 Tower kernel: nvme nvme0: Abort status: 0x0
Nov  3 02:02:11 Tower kernel: nvme nvme0: I/O 140 QID 8 timeout, reset controller
Nov  3 02:02:41 Tower kernel: unregister_netdevice: waiting for lo to become free. Usage count = 1
Nov  3 02:03:12 Tower kernel: nvme nvme0: I/O 214 QID 0 timeout, reset controller
Nov  3 02:03:35 Tower kernel: nvme nvme0: Device not ready; aborting reset
Nov  3 02:03:35 Tower kernel: nvme nvme0: completing aborted command with status: 0007
Nov  3 02:03:35 Tower kernel: blk_update_request: I/O error, dev nvme0n1, sector 263207480

 

That NVMe device uses a Marvell controller, so it could be that, but I assume you're using a PCIe adapter, so can also be that or the device itself, try the Samsung to see if it's any different.

Yea im useing a adapter card. could it be the card? when i bought the Samsung i bought another adapter card. (the new card had some good reviews on amazon) i might try and swap the adapters? also i swaped the nvme drive's pcie slots so i dont think its the MOBO but dont know. what do you think? also i red the reviews on the wd drive some people said the drived failed like 30 days in. what you think

question is there a script to save the log and the diagnostics log like the backup plug in to the array or usb key befor shutdown

Edited November 5, 20178 yr by Gary489

5 hours ago, Gary489 said:

question is there a script to save the log and the diagnostics log like the backup plug in to the array or usb key befor shutdown

Tips & tweaks plugin if it's for a normal shutdown.   Fix common problems plugin in troubleshooting mode for random crashes

8 hours ago, Gary489 said:

also i swaped the nvme drive's pcie slots so i dont think its the MOBO but dont know. what do you think?

 

It could be the board/slot, the adapter or the NVMe device, you need to start ruling them out, using a different adapter, etc.

I have the same exact disk, attached straight to the Gigabyte GA-H270N-WIFI motherboard, and this morning when I woke up I had a notification saying that my Cache disk was missing. It had a green ball next to it on the Main page, but when I went into its page it said that the drive was "spun down", and when I hit the Spin Up button, nothing happened.

 

Then I did a reboot and nothing happened.

 

Then I did a power down, and then power up again, and now it's back.

 

But seeing as you had the same exact issue with this exact drive I'm bailing on it. Gonna run the mover and hope it gets my Plex install out of there, so I don't have to refresh all metadata etc.

 

I ran diagnostics but all the SMART report says for it is: "Rea d NVMe SMART/Health Information failed: NVMe Status 0x4002"

tower-diagnostics-20171205-1024.zip

9 minutes ago, nadbmal said:

"Rea d NVMe SMART/Health Information failed: NVMe Status 0x4002"

Thta's normal with v6.3.5, NVMe SMART support is better on v6.4.

 

Those NVMe devices use a Marvell controller, probably not the best option for unRAID with the known issues with Marvell controllers.

7 minutes ago, johnnie.black said:

Thta's normal with v6.3.5, NVMe SMART support is better on v6.4.

 

Those NVMe devices use a Marvell controller, probably not the best option for unRAID with the known issues with Marvell controllers.

Yeah I just don't trust it anymore.

Since it has a 5 year warranty, I'm gonna see if I can return it and get my money back and go for a Samsung 960.

I ditched the wd black drive and replaced it with a Samsung 960. The samsung drive is way better. And its even faster then the wd black. 

2 hours ago, nadbmal said:

I have the same exact disk, attached straight to the Gigabyte GA-H270N-WIFI motherboard, and this morning when I woke up I had a notification saying that my Cache disk was missing. It had a green ball next to it on the Main page, but when I went into its page it said that the drive was "spun down", and when I hit the Spin Up button, nothing happened.

 

Then I did a reboot and nothing happened.

 

Then I did a power down, and then power up again, and now it's back.

 

But seeing as you had the same exact issue with this exact drive I'm bailing on it. Gonna run the mover and hope it gets my Plex install out of there, so I don't have to refresh all metadata etc.

 

I ran diagnostics but all the SMART report says for it is: "Rea d NVMe SMART/Health Information failed: NVMe Status 0x4002"

tower-diagnostics-20171205-1024.zip

Get a samsing drive it way better