Hi, everyone!
I'm running a small home server (see signature) with 4 2TB SATA-drives as a single RAIDZ1-vdev in a single pool.
The drives were bought more than four years ago, are Samsung SpinPoint drives (desktop drives, not 24/7 ready) because back then I didn't know any better.
Last weekend, one drive stopped responding, was detatched from the bus and the vdev was degraded. Since its a RAIDZ1, it continued working.
I took a backup of the most important data (which I do every two weeks or so), so loss of the pool would be annoying, but nothing really important would be lost. Then I shut down the server, checked the cabling, wrote down the serial numbers and locations of the disks and booted the server up again. And ... the drive started working again. I did a full scrub, which corrected a lot of errors on the drive (of course, since the data on the other drives changed, while it was dead), but they were "only" checksum errors, not read or write errors.
When the scrub finished, I did a long SMART-selftest on all of the drives and it finished without an error. BUT, looking at the smart-data of the previously-failed drive, it starts producing recoerable Read-Errors. So, I figured, the drives are near the end of their useful lives. And sonst all four drives are exactly the same size and have run in the same NAS (and now server), they will all fail within the near future. So, I ordered four WD Red 4TB, which will arrive by tomorrow.
Reading the manual, i saw that I can insert a new drive (as a 5th drive) and replace an old one (replace-button). When that finishes, I can shut down the server and physically remove the old drive. I'll do this four times and end up with four new drives with hopefully intact data (because I never lose redundancy). I know, there's some risk involved, but as I said, loss of the pool would be annoying, but not catastrophic. And I don't have the space to backup everything.
Now my question (really long introduction, sorry for that):
I have 6 SATA ports on my mainboard (which is good), four of them being SATA2 and two being SATA3. During the changing of drives, I will have to switch the SATA ports once. Is that a problem?
In short: If i Shut down the server, switch SATA ports and boot it back up, will the pool still be operational? I guess so, because FreeNAS uses UUIDs and such, but I just wanted to ask.
Thank you!
Bernhard
I'm running a small home server (see signature) with 4 2TB SATA-drives as a single RAIDZ1-vdev in a single pool.
The drives were bought more than four years ago, are Samsung SpinPoint drives (desktop drives, not 24/7 ready) because back then I didn't know any better.
Last weekend, one drive stopped responding, was detatched from the bus and the vdev was degraded. Since its a RAIDZ1, it continued working.
I took a backup of the most important data (which I do every two weeks or so), so loss of the pool would be annoying, but nothing really important would be lost. Then I shut down the server, checked the cabling, wrote down the serial numbers and locations of the disks and booted the server up again. And ... the drive started working again. I did a full scrub, which corrected a lot of errors on the drive (of course, since the data on the other drives changed, while it was dead), but they were "only" checksum errors, not read or write errors.
When the scrub finished, I did a long SMART-selftest on all of the drives and it finished without an error. BUT, looking at the smart-data of the previously-failed drive, it starts producing recoerable Read-Errors. So, I figured, the drives are near the end of their useful lives. And sonst all four drives are exactly the same size and have run in the same NAS (and now server), they will all fail within the near future. So, I ordered four WD Red 4TB, which will arrive by tomorrow.
Reading the manual, i saw that I can insert a new drive (as a 5th drive) and replace an old one (replace-button). When that finishes, I can shut down the server and physically remove the old drive. I'll do this four times and end up with four new drives with hopefully intact data (because I never lose redundancy). I know, there's some risk involved, but as I said, loss of the pool would be annoying, but not catastrophic. And I don't have the space to backup everything.
Now my question (really long introduction, sorry for that):
I have 6 SATA ports on my mainboard (which is good), four of them being SATA2 and two being SATA3. During the changing of drives, I will have to switch the SATA ports once. Is that a problem?
In short: If i Shut down the server, switch SATA ports and boot it back up, will the pool still be operational? I guess so, because FreeNAS uses UUIDs and such, but I just wanted to ask.
Thank you!
Bernhard
