Cache SSD drive disappears when inserting disk in removable bay

[bclinton]

Greetings folks. I recently built my unraid server and it is humming along. I decided to start adding some removable bays to it. Here is my setup. Dell H310 connected to 1 partity 8tb and 2 8tb data disks as well as a hot swap drive bay I just added. 2 1TB SSD drives connected to the MB sata ports (cache). 1 External USB backup drive 8TB named offsite. When I take the offsite drive and insert it into the hot swap drive bay one of my cache SSD drives disappears and I get a warning. Move the offsite drive back to USB and reboot and it all returns to normal. I can take another 8TB blank and insert it into the hot swap bay and it is recognized and the SSD drives are fine. Not sure why this is happening with the offsite drive. I have a plus license so that is not the issue. I am going to rack mount my server and wanted to use the hot swap bays for my offsite backups instead of the USB drive.

Edited March 14, 2021 by bclinton

[JorgeB]

Please post the diagnostics: Tools -> Diagnostics

[bclinton]

2 hours ago, JorgeB said:

Please post the diagnostics: Tools -> Diagnostics

 

Thanks for assisting. Please see attached. 3-12 and 3-13 were the 2 days that I replicated the problem several times. As a side note.....when this first occurred I could have sworn that I set my 2 SSD drives up in a a BTRFS raid 1 setup and it was 1 TB with the 2 TB drives in the pool. During all of this I see that I now have 2 TB capacity with my cache pool which indicates they are not mirrored. I started the array at least once early on when it showed one of the SSD drives as missing. Not sure if that could cause an issue with the cache pool changing it's format. 

unraid-diagnostics-20210315-0609.zip

Edited March 15, 2021 by bclinton

[JorgeB]

I'm not seeing any issues with a cache device going offline in the time covered by the syslog, it is using single profile though, and it shows some previous data corruption in one of the devices, you should run a scrub, some more info here.

 

After fixing the pool reboot, reproduce the problem and post new diags.

[bclinton]

24 minutes ago, JorgeB said:

I'm not seeing any issues with a cache device going offline in the time covered by the syslog, it is using single profile though, and it shows some previous data corruption in one of the devices, you should run a scrub, some more info here.

 

After fixing the pool reboot, reproduce the problem and post new diags.

 

Thanks. I have a parity check running now and mover so it will be later in the day. I will report back. I am reading up on the thread you provided. Good info. Am I assuming correctly that running btrfs scrub in the terminal is what you mean? Are there risks with losing data when I do that? I will have a good back but wanted to know what to expect when I do that.

 

Thanks again for lending a hand!

[JorgeB]

16 minutes ago, bclinton said:

Am I assuming correctly that running btrfs scrub in the terminal is what you mean?

Terminal or GUI.

 

17 minutes ago, bclinton said:

Are there risks with losing data when I do that?

No.

[bclinton]

On 3/15/2021 at 8:28 AM, JorgeB said:

Terminal or GUI.

 

No.

 

Greetings.....additional info and the new diag files. I installed a second removable tray to the server. When I put offsite drive in the new removable bay the problem occurs. I took the offsite drive and inserted it into the original hot swap bay and the problem occurred. I move the offsite drive back to the USB dock and it returns to normal.  What is weird is I can take another drive and insert it into that same bay and no problem. 

 

unraid-diagnostics-20210316-0903.zip

Edited March 16, 2021 by bclinton

[JorgeB]

I see ATA1 link going down and getting disable:

 

Mar 16 08:38:17 unraid kernel: ata1: SATA link down (SStatus 0 SControl 300)
### [PREVIOUS LINE REPEATED 2 TIMES] ###
Mar 16 08:38:28 unraid kernel: ata1.00: disabled

 

If that it's happening when you hot unplug a drive it's likely a problem with the backplane, or that slot in the backplane, or it doesn't really support hot-swap, best to avoid it.

[bclinton]

53 minutes ago, JorgeB said:

I see ATA1 link going down and getting disable:

 

Mar 16 08:38:17 unraid kernel: ata1: SATA link down (SStatus 0 SControl 300)
### [PREVIOUS LINE REPEATED 2 TIMES] ###
Mar 16 08:38:28 unraid kernel: ata1.00: disabled

 

If that it's happening when you hot unplug a drive it's likely a problem with the backplane, or that slot in the backplane, or it doesn't really support hot-swap, best to avoid it.

 

Thanks for the info. I am still wondering why I can insert another drive into the same drive bays and all is normal. It only seems to happen with that one drive. I'll just keep using the USB cradle for my offsite backups which is fine. Weird but at least I have more info on it. Thanks!