Keep getting IO failure on usb replication drive

[Adcam]

Hello,​

I keep getting this error repeatedly. I have gone 3 months without a single IO error for the drive. Its configured properly as a replication drive to backup the main drive pool. It seems the only way I can get rid of this is by completely removing the replication pool, reformatting the replication drive and starting from scratch. Then I am good for a few months then I get these errors again. I'm pretty new to all this so pardon my lack of knowledge. I have looked into this in many other forums but I am stuck. There has to be an easier way to fix this without having to redo the entire replication drive setup. Any help would be appreciated!

​

Pool Replication Drive state is UNAVAIL: One or more devices are faulted in response to IO failures.
The following devices are not healthy:

• Disk Seagate FireCuda HDD Hub 00000000NAD415MW is REMOVED

 

[Arwen]

There are known problems with SMR disks, which the Seagate FireCuda HDD appears to be. Topping that off with USB may make problems happen more often.

One thing to try, (unless you are doing it already), is to power down the USB drive every now and then. This resets the drive's firmware and the host server's driver.

ZFS and always powered on USB drives can be a problem, though low use pools like USB boot pools, seem to work okay.

 

[Adcam]

There are known problems with SMR disks, which the Seagate FireCuda HDD appears to be. Topping that off with USB may make problems happen more often.

One thing to try, (unless you are doing it already), is to power down the USB drive every now and then. This resets the drive's firmware and the host server's driver.

ZFS and always powered on USB drives can be a problem, though low use pools like USB boot pools, seem to work okay.

I do power cycle the drive every so often. It is powered on all the time and I know USB IO errors are a constant problem. The system seems to go through and disable all ports on the server. I also cant do a clean restart when I get this error. It gets stuck on unable to terminate processes and I have to power cycle the server. Would a PCI USB card do the same thing?

 

[Arwen]

...
Would a PCI USB card do the same thing?

Don't know.

Always on USB drives, with always imported ZFS pools are problematic. USB attached data drives are not recommended.

I know that is not the answer you want to hear, but problems like this are hard to troubleshoot. Many of us in the forums are conservative, so anything that reduces reliability of our data tends to be removed from our NASes. Meaning we don't troubleshoot such issues, we remove them.

Don't forget you have a SMR drive to top off the problems.

 

[Adcam]

Don't know.

Always on USB drives, with always imported ZFS pools are problematic. USB attached data drives are not recommended.

I know that is not the answer you want to hear, but problems like this are hard to troubleshoot. Many of us in the forums are conservative, so anything that reduces reliability of our data tends to be removed from our NASes. Meaning we don't troubleshoot such issues, we remove them.

Don't forget you have a SMR drive to top off the problems.

Thank you for that info. What would you recommend as a replication device?

 

[Arwen]

Sorry, I don't have any recommendations, except not an SMR disk.

 

[Adcam]

Sorry, I don't have any recommendations, except not an SMR disk.

Okay. Thanks for your help.