Unraid kernel: I/O error

[GaBoNtZ]

Hi all,

 

Recently I replaced my mirrored cache drives with 2 SSD enclosures in which I have 2 NVME drives in each, set up in SPAN.

Since I replaced them, I am getting the I/O error in the syslog.

Unfortunately I can't download the diagnostics, as it stays like that forever... Video here.

I do have access to syslog as it's saving it on a usb stick.

I should mention I forgot to preclear the drives.

Couldn't alocate both at the same time to the cache pool, so I left an old one and added a new one and the system imediately started to mirror them.

Might that be my issue here?

 

TIA

 

syslog-1703604119

Edited December 27, 2023 by GaBoNtZ

[JorgeB]

2 hours ago, GaBoNtZ said:

with 2 SSD enclosures in which I have 2 NVME drives in each

Are these using USB?

 

Log shows errors for sdg and sdj, are these the pool devices?

[GaBoNtZ]

33 minutes ago, JorgeB said:

Are these using USB?

 

Log shows errors for sdg and sdj, are these the pool devices?

Enclosures are SATA. these.

And yes, those are the drives.

Edited December 27, 2023 by GaBoNtZ

[JorgeB]

The link is not working, so are the devices SATA, not NVMe?

[GaBoNtZ]

4 minutes ago, JorgeB said:

The link is not working, so are the devices SATA, not NVMe?

SATA enclosures for NVME with RAID options.

Maybe this time the link works here.

There are 2 NVME devices in each, with RAID set to SPAN.

The cache pool is using both adaptors/drives, and are mirrored.

Edited December 27, 2023 by GaBoNtZ

[JorgeB]

According to the link it's for M.2 SATA, not NVMe, in any case I would recommend retesting without the enclosure, not sure it will be supported.

[GaBoNtZ]

16 minutes ago, JorgeB said:

According to the link it's for M.2 SATA, not NVMe, in any case I would recommend retesting without the enclosure, not sure it will be supported.

The drives are these. And you are right, M.2

It works, but I just noticed those errors in the live syslog.

Edited December 27, 2023 by GaBoNtZ

[JorgeB]

Some of the errors may be from lack of TRIM support, still would recommend testing without the enclosure, Unraid should have direct access to all devices.

[GaBoNtZ]

1 hour ago, JorgeB said:

Some of the errors may be from lack of TRIM support, still would recommend testing without the enclosure, Unraid should have direct access to all devices.

Thanks for your time. Motherboard can't host that many M.2's.

I'm just moving the data off of them and do a proper initialization.

Will test again and if still the same, I think I'll just get 2 normal SSD's.

I have those drivers and I thought to make use of them.

 

Thank you.

[GaBoNtZ]

So mover will not move everything from that pool.

The pool is a cache pool for appdata, isos, system and domains and all of them were set like in the prhoto.

It doesn't seem to run anymore, because I can't see any new read/writes on any of the pools, but the move button on the main is greyed out, saying mover running.

It's been like this for many hour.

Anyway, question is: if it didn't move everything, can I move manually, format the drives and move them back?

Also, I don't understand why some of the folders show up like this:

In my understanding, that means those folders are still on the cache drive.

Docker and VM are stopped and not running.

 

[JorgeB]

Enable the mover logging, run the mover, post new diags.

[GaBoNtZ]

Mover loggings are enabled, but can't download the diags.

It just hungs for hours doing something that will never finish.

[JorgeB]

Try to get the syslog:

cp /var/log/syslog /boot/syslog.txt

 

[itimpi]

Had you disabled the docker and VM services as is mentioned here for doing that in the online documentation accessible via the ‘Manual’ link at the bottom of the GUI or the DOCS link at the top of each forum page. 

[GaBoNtZ]

1 hour ago, itimpi said:

Had you disabled the docker and VM services as is mentioned here for doing that in the online documentation accessible via the ‘Manual’ link at the bottom of the GUI or the DOCS link at the top of each forum page. 

Yes, both set to "No"

[GaBoNtZ]

1 hour ago, JorgeB said:

Try to get the syslog:

cp /var/log/syslog /boot/syslog.txt

 

syslog.2syslog

Edited December 28, 2023 by GaBoNtZ

[JorgeB]

Log is filled with ATA errors for sdj, try to manually copy what you can, mover probably not going to work.

[GaBoNtZ]

10 minutes ago, JorgeB said:

Log is filled with ATA errors for sdj, try to manually copy what you can, mover probably not going to work.

That's what I'm trying to do now. TA

[GaBoNtZ]

I'm just preclearing the drives.

I'm guessing I'll need normal SSd's with these speeds.

 

#################################################################################################### # Unraid Server Preclear of disk _1T_dWHWa6 # # Cycle 1 of 1, partition start on sector 64. # # # # Step 1 of 3 - Zeroing in progress: (79% Done) # # # # # # # # # # # # ** Time elapsed: 0:39:16 | Write speed: 325 MB/s | Average speed: 325 MB/s # # # #################################################################################################### # Cycle elapsed