Replacing 2TB harddrives for 4TB harddrives, one by one?

[Norlig]

Hi,

I'm a bit confused regarding how I should go ahead upgrading my harddrives.
I have 4x2TB Harddrives at the moment, in Raid Z, its not sensitive data, so Im not that concerned about it.

I plan to replace the harddrives with 4x4TB Drives and was wondering if I can do it by unplugging one, putting the 4TB in and letting the lost drive get recovered from the other 3, or if I can plug the new drive in as an additional one and copy the files from one drive at a time (a coworker said to do this)

Sorry if this is all documented somewhere, I have tried looking for it.
If someone could explain or link me the correct thread, I would appreciate it highly!

 

[danb35]

If you have a spare SATA port on your motherboard and space in your case for another drive, here's the safest way to do it:

• Shut down the server, install one of the 4TB drives, power back on.
• Log in to the web GUI
• Go to Storage -> select your volume -> Volume Status (the button at the bottom of the page that looks like a sheet of notebook paper)
• Select one of your 2TB disks from the list and click the Replace button at the bottom of the page
• In the pop-up, select your new 4TB disk from the drop-down and click Replace Disk.
• Wait for the resilver to finish (you can monitor it from the CLI by running 'zpool status', or from the volume status page.
• Power down the server, remove the 2TB disk you just replaced (make sure it's the right one--match up the serial number using the View Disks page), and return to step 1.

If you don't have room for a spare disk, the process is a little different, and a little more dangerous because you won't have any redundancy while the resilvers are happening. You'll log in to the web GUI, go to Storage -> select your volume -> Volume Status, select one of your disks, and hit the Offline button. Then go to step 1 above.

This is very similar to the procedure in the manual for replacing a failed disk, but not identical.

 

[Norlig]

Thank you, seems pretty straight forward :)
I do have 2 spare sata ports on the motherboard, so I can do the safe approach.

Its also my understanding that untill I have replaced all 4, they will all show as 2TB?
they will show as 4TB when all 4 has been replaced?

 

[Z300M]

If you have a spare SATA port on your motherboard and space in your case for another drive, here's the safest way to do it:

• Shut down the server, install one of the 4TB drives, power back on.
• Log in to the web GUI
• Go to Storage -> select your volume -> Volume Status (the button at the bottom of the page that looks like a sheet of notebook paper)
• Select one of your 2TB disks from the list and click the Replace button at the bottom of the page
• In the pop-up, select your new 4TB disk from the drop-down and click Replace Disk.
• Wait for the resilver to finish (you can monitor it from the CLI by running 'zpool status', or from the volume status page.
• Power down the server, remove the 2TB disk you just replaced (make sure it's the right one--match up the serial number using the View Disks page), and return to step 1.

If you don't have room for a spare disk, the process is a little different, and a little more dangerous because you won't have any redundancy while the resilvers are happening. You'll log in to the web GUI, go to Storage -> select your volume -> Volume Status, select one of your disks, and hit the Offline button. Then go to step 1 above.

This is very similar to the procedure in the manual for replacing a failed disk, but not identical.

So when both drives are connected at the same time, the vdev still has redundancy while the resilvering is in progress?

 

[GrumpyBear]

Thank you, seems pretty straight forward :)
I do have 2 spare sata ports on the motherboard, so I can do the safe approach.

Its also my understanding that untill I have replaced all 4, they will all show as 2TB?
they will show as 4TB when all 4 has been replaced?

Here is the relevant section from the documentation for replacing disks to grow capacity. Here is the section on replacing a failed drive. Make sure you verify that autoexpand is enabled.

You might want to consider doing a backup and restore and create a RAIDZ2 to host your data on. You will lose capacity (8TB versus 12TB) but you will get a lot more fault tolerance. Or look at 6 3TB in RAIDZ2 (same capacity) or the really inexpensive approach by adding 2 4TB disks to create an array with your existing 4 2TB disks (6TB -> 8TB initially) then replacing the 2TB disks as you can afford additional 4TB disks.

Note that irrespective of the approach you take you should burn-in the new disks in to get rid of any "infant mortality" issues and give you confidence in them before you replace them. There is a post here with some details. You can do this on one of your spare SATA interfaces while the system is running.

 

[SirMaster]

So when both drives are connected at the same time, the vdev still has redundancy while the resilvering is in progress?

Yes, the zpool is still using the old drive until the resilver completes successfully. Only then will it kick out the old drive.

 

[Norlig]

Resilvering right now, tested the discs yesterday and they seem to hold for now, but will also replace them by one disc per week.

What I am wondering now, is if I need to keep the discs in the same SATA slot after resilvering an disc?

Can I Resilver Disc #1 in Sata port 1, with Disc #5 in sata port 5, then move Disc #5 to sata port 1 after the resilver?
Or will that cause issues/need configuring?

 

[joeschmuck]

The SATA ports you use are not relevant with respect to FreeNAS so you do not need to move them around but you can if you want to, just shut it down before rearanging the ports.