Replacing a disk in raid-z2

I have a 6-drive raid-z2, and one of the disks has errors and needs to be replaced
The disks were purchased online, and I have to to send them back by next week to get a refund
If I order a new disk it won’t arrive before I ship the disk
Is it ok to keep using the system for a couple of weeks in degraded mode (a missing disk), or should I purchase a disk right awaybfrom local stores (much more expensive)?

It would be OK-ish to work for a few days in degraded mode. But it is better to buy an extra disk right now, possibly even resilvering with the partially defective drive still in before sending the drive for RMA. (You do not have the time to fully burn-in the new drive AND then to resilver by next week though.)

Then keep the seventh drive at hand as a pre-validated cold spare.

1 Like

please explain the resilvering process you are describing. If I take out (offline) the defective drive, I’ll be left with a 5-disk raid in a degraded 6-drive raid-z2. What resilvering are you proposing in that stage?
Additionally, when you say that is “OK-ish for a few days”, is there quantitative data to tell me what are the risks?
If I understand correctly, if another disk dies during the wait period, I could turn off the server, waiting for the replacement disks, without losing any data. Is that correct?

Since this is raidz2 you could even run it with two failed drives… but any further issue would mean lost (or at least unretrievable) data.
Rather, one should aim to have full redundancy (-cies) at all times—and repair damage as soon as possible.

Which is why the best course of action, budget allowing, would be to get a drive as soon as possible. Plug it if you have a spare SATA port; otherwise, get a USB adapter, and put the failing drive on it to free its port. Then, with all seven drives attached, replace the failing drive from the GUI—that is, resilver with full redundancy, without offlining the failing drive.

1 Like

In the past, I have replaced disks by offlining them and then replacing the drive (starting a lengthy resilvering process)
Why do you prefer to perform the resilvering with all drives online? (creating a temporary 7-disk raid-z2, and then offlining a disk)

Is that worthwhile even if I have to get a USB-SATA adapter?

It provides increased resiliency. The drive you are replacing may still be contributing resiliency to your pool. Offlining it means you have now sufferered a full drive loss. In a RaidZ1 or 2-way mirror pool any further single bit errors will result in data loss. RaidZ2 is not as critical.

But why increase the risk of data loss when you don’t have to?

The problem is that if the drive is failing in a way that it’s very slow to read, then it may be better to offline it, but you can do this after the replace if it is going very slowly, and secondly, you do require a spare bay.

1 Like

I’d argue to get a powered sata adapter if at all possible; one that doesn’t just rely on usb for power. Can reduce flakiness.