Hi.
TLDR: Is there a way to “stop” a pool to protect it from underlaying hardware issues? (the equivalent of mdadm stop/start)
When I export the pool, if I understand correctly - it’s still there, I still see it via cli so I assume that exporting is not enough?
Longer version:
I am going to do some re-arrangements of my hardware that include changing SAS cables, etc…
My first attempt of using my JBOD resulted in tons of errors that even left a few permanent UDMA SMART errors on one of my drives. Re-attaching the cable and using a different SAS port has cleared it all and it has been (touch wood) pretty solid for the last few weeks.
I am afraid that I might not get it 100% working at my first attempt, I don’t want ZFS to mark any of my drives as faulty, especially when it can happen to all of them so I am looking for a way to put them aside while I experiment with a few other drives before I put them back in.
What is the best method to stop them? (ideally - I want them completely outside the JBOD while I experiment with it) and how do I re-attach them? In the old linuxraid days I can simply stop the array with mdadm and then start back again when I am good for it, what is the equivalent here?
That is more or less what ZFS export / import does. When a ZFS pool is exported, nothing happens to it by ZFS. The OS on the other hand can still see the pool’s disks, because, well, they are connected. Just like Linux MD-RAID devices that are stopped, the OS can still see the underlying disks.
Unless you have other disks to sub in for the ZFS pool disks, you are somewhat constrained in your testing. Plus, if you use TrueNAS for the testing, it would normally import the ZFS pool as part of it’s boot procedure.
I have other drives to use to test this so I am not too worried.
What happens if TrueNAS starts and ALL drives for a pool are gone? Can I plug them in afterwards , export and then import to TrueNAS? What would be the procedure (I don’t want to lose the data for doing something silly. I assume some resiliency but not total resiliency )
During a recent troubleshooting i removed all drives and restarted truenas. Upon reboot it complained that all drives were missing(what was to be expected) but after shutting down the machin, reconnecting the drives and starting the box, truenas picked them up as if nothing had happened.
II don’t have a backup, this data is not important enough to back it up and I can always redownload it, important stuff is backed up to 2 separate clouds in addition to the original machine.
But to redownload everything from the internet would be a PITA, so I’d rather avoid that
)