I have a two hard drive mirror connected to a Zimablade through its two native SATA ports.
I issued a shutdown command from the UI. The Zimablade powered of within a few seconds. When I powered it back on truenas reported checksum errors on one of the drives. I ran a scub which found no errors to repair so I’m guessing auto healing took care of the errors.
It seems that truenas powered off the Zimablade before some cache writing had completed. If that’s correct how can I fix that? Even injecting some delay would be sufficient but I don’t know how to do that. Thoughts?
Is this issue reproducible or has this only happened once? ZFS wouldn’t have shut down before the writes were marked as complete
I would double check the data and power connections while the system is off. Running and checking SMART Long tests on the data drives might be a good idea.
Track the drives by the serial numbers as the drive position names can changes upon booting. SDA could be SDB the next boot.
I hadn’t tried a second time. Now I have and there were no errors on startup this time. Ran SMART long test which reported nothing so I’ll just keep monitoring and see what happens.
That’s not possible in my current setup since the Zimablade only exposes one power connector for it’s two SATA ports but since it’s their cable and the one that comes with their NAS setup I’d have to assume power is OK. But your point is well taken and if the problem recurs I’ll figure out a way to check on the power. Eventually I will be installing a PCIe SATA card to allow for more drives. Then I will have to get a separate power supply but that is for another day.
OK. So if and when I add ports I want a PCI card that does not do so through port multiplying even if that means less ports. Nor would I want to implement RAID in hardware (I think I knew that one). For example cards using ASMedia ASM1166 or ASM1164. Do I have that right?