Bug replication test, please

[diedrichg]

Would one of you with a test system try out this possible bug before I submit it, please. I'm on 9.2.1.9

1. Build SMART test and Scrub schedules on your volume. (I'm in RAIDZ2 with four drives.)
2. Shut down.
3. Swap SATA cables between two of the drives leaving the cables on the same ports on the motherboard.
4. Boot.
5. Check your SMART test and Scrub schedules.
6. Bug?: Are only two of the four drives in the pool being included in the test schedules? Two of my drives were not included and it has been several months since I swapped those cables!

 

[Robert Trevellyan]

Two of my drives were not included and it has been several months since I swapped those cables!

FYI, when you replace a drive you have to add the new drive to the SMART tests, because FreeNAS tracks drives by unique ID, not by device name. Could this explain your observation?

 

[Ericloewe]

Would one of you with a test system try out this possible bug before I submit it, please. I'm on 9.2.1.9

1. Build SMART test and Scrub schedules on your volume. (I'm in RAIDZ2 with four drives.)
2. Shut down.
3. Swap SATA cables between two of the drives leaving the cables on the same ports on the motherboard.
4. Boot.
5. Check your SMART test and Scrub schedules.
6. Bug?: Are only two of the four drives in the pool being included in the test schedules? Two of my drives were not included and it has been several months since I swapped those cables!

Cannot reproduce.

When I cleaned/upgraded my server in February, the drives went back in in a random order. No problems, all schedules continued running normally.

9.3, though, but I don't see how moving drives around would cause trouble, since they're referenced by GPTID.

 

[diedrichg]

Could have very well been user error when initially creating the schedules. Oh well, thanks for testing, maybe I'll try this again the next time I open it up.