Multipathing After TrueNAS Scale Upgrade

TrueNAS General
,
1

I recently went through a multi-step upgrade from TrueNAS Core 13.0-U6.7 → Scale Dragonfish → ElectricEel, and while I did my best to review the release notes, I completely overlooked that my NAS is set up with multipathing to a set of 15 SAS drives.

My Setup:

• Drives: 15 SAS disks in an old EMC VNX DAE

• Multipath: Both SAS connections from a single controller going into the disk tray

• Current State:

• Everything seems fine – the storage pool is happy and healthy, and all disks are visible within the Disks screen.

• Issue: The Disks screen shows 14/15 disks as “N/A” under the Pool column, while the last one shows correctly assigned.

• Pool Management: When I check under Pool → Manage Devices, I can see my two VDEVs, and all disks are properly assigned under each RAIDZ2.

What I’ve Found So Far:

As I dig into this more, it looks like multipathing was never recommended in TrueNAS and was removed at some point after I originally set it up. On top of that, it’s not even configurable in Scale. According to the documentation, I shouldn’t have been able to complete the upgrade with multipathing still in place but here we are.

I want to break down the multipath setup cleanly without risking data loss. My current idea is:

  1. Take the Pool offline

  2. Break the multipath configuration

  3. Re-import the pool

Would this be the best approach, or is there a safer/easier way to go about it?

2

Generally this is the case.

How do you plan to break down multipath config… just changes in hardware layout?

Do you have a backup[ of the valuable data?

3

Yes, physically remove one of the multipath cables.

I don’t have a backup of the data, would be painful but it’s replaceable.

4

I went ahead and tackled this today, and… it didn’t go quite as planned.

I took the pool offline, broke the multipath setup, and attempted to re-import the pool. The mistake I made was probably moving a bit too quickly. I started getting I/O errors everywhere, the pool failed to import, and eventually showed as suspended—definitely a few “oh crap” moments.

What ended up fixing things was:

• Shutting down the TrueNAS server

• Power cycling the disk tray

• Then powering the server back on

Thankfully, everything came back happy after that.

My Updated Recommendation

For anyone in a similar setup, here’s what I’d suggest instead:

  1. Take the pool offline

  2. Shut down the TrueNAS server

  3. Break the multipath configuration

  4. Power everything back on

  5. Re-import the pool

That should give your hardware a clean state and avoid the I/O errors I ran into.

Hope this helps someone avoid the same stress!

1 Like