What’s the most optimal way to “shift” data around the local system. By that I mean I currently have an 8 x 8Tb disk raidz1, I want to replace those 8 disks. So I’ve inserted some temporary 18Tb disks that I want to transfer the data to, then pull the 8 disks and replace them with 24Tb disks (going to smaller vdevs or raidz2). My question is, what’s the most optimal way to play this little dance? I don’t care about having to recreate pools, configs etc, just would prefer not to lose the data.
Initial thought was setup a replication task, let that keep the data replicated, once I have the new 24Tb disks in hand, I pull the 8Tb disks (destroying that pool…not looking to do it “gracefully”) then restore to the new pool. Is this a reasonable approach, is there a better way, would a simple rsync job be better?
Replication is how I’d handle it as well, but I’d offline the 8x8TB pool politely, by just choosing the “Export Pool” option without checking “Destroy Data.” This will give TrueNAS the chance to relocate the system dataset and other log/running pieces without having to do it suddenly.
Then insert your N x 24TB (however wide you’re running it) disks, replicate back if you choose, and politely offline the 18TB pool - unless you’re keeping that one around as well.
Thanks! That makes sense to offline them. By graceful I really meant I didn’t want to try and do it one disk at a time and once all replaced expand or something like that. Will definitely do it via offline/export.
On the restore, (having never restored before) is it another replication task to repliacte it back or use the restore function on the replication job?
Basically you’ll be creating a new replication job, because the initial will be from 8TB-RAIDZ1 to 16TB-temporary pool - and then that job will be stopped, the 8TB-RAIDZ1 pool will be taken offline, and then you’ll make a new job to go from temporary-16TB to shiny-new-24TB-RAIDZ2
