Multipath multipath/disk4 connection is not optimal

[tensi0n]

Greetings everyone!

I am a bit of a novice when it comes to errors such as the one in the title of this post. I've been using FreeNAS/TrueNAS for years with little to no issues. Hence why I am a novice....ie I rarely have to troubleshoot my nas! :) Anyways...I'm not sure exactly what I should do about this error that I am getting. So far I have shutdown the nas, re-seated disk4, booted everything back up, and I still get the error. I've read over some of the existing posts talking about this issue and it appears I may need to replace the disk. I'm getting no other disk issues and overall the pool is reporting healthy (green check mark). Any assistance would be greatly appreciated!

Version:
TrueNAS-12.0-U2.1

 

[Evertb1]

As suggested (well a bit stronger then that) in the Forum rules you should tell us how your system is configured and what hardware is used. Especialy with an problem like this that would be of great help.

 

[tensi0n]

Sorry about that. I just read the instructions and my fault for not paying attention. The following is my hardware:

TrueNAS is running on a Dell PowerEdge R310. The R310 is then connected to a SuperMicro 4U JBOD (CSE-PTJBOD-CB2)
via a LSI 9200-8e 6Gbps 8-lane external SAS HBA (flash to IT mode) controller.

CPU: Intel(R) Xeon(R) CPU X3450 @ 2.67GHz
Memory: 32GB
All the drives (including the one giving the issue): Seagate ST4000NM00023 4TB 7.2K 6G/sec SAS 3.5"
I just have one pool set up with 16 of the above mentioned drives.

This is a home lab environment and not production.

Thank you.

 

[Michael Phelps]

Hello,

Did you ever figure out what was causing the 'connection is not optimal' error? I am getting a similar error.

 

[tensi0n]

Not really... I ended up powering down the nas and swapping the location of the drive with a one of the other drives. I rebooted and the problem hasn't come back. I read somewhere on reddit someone saying they did that and it fixed the issue. You may even try replacing the drive. Hope that helps.

 

[Evertb1]

As far as I know Multipath I/O (multipath connection) is only in play if you have managed, multiple physical connections to a physical or logical block storage device. But it should be something that is configured. If your system is not configured that way (if your hadrware even support it) you should not have this problem. So while I have no idea why you had that problem (I have no meaningfull experience on this subject) I suspect it was a glitch in your system. Maybe caused by the combination of your jbod hardware an the LSI HBA. It would not surprise me if one of those or both can be configured for Multipath I/O.

One of these days I really should ask one of our storage specialists if he or she can shine a light on this problem.