Replacing failing but not failed hard drive

[RonRN18]

In one of my volumes, I have four 3TB drives in a RAIDZ1 configuration. I've already had to replace one of the drives a year or so ago when it failed. I now have another drive that is failing but has not yet completely failed. I've been having sector issues but otherwise working without decreased performance yet. I could have sworn I had read that there was a way to add the new replacement drive to the system and have it essentially copy the contents of the failing drive to the new drive without going to a degraded status and resilvering that occurs when the drive is actually failed. I have been trying to find the thread I thought I read but I'm not finding it after using many search strings. Am I imagining that there was a way to do this? If not, can someone point me to the process, as I'm not finding it in my searches?

 

[Chris Moore]

Hard Drive Troubleshooting Guide (All Versions of FreeNAS)
Basic information on how to troubleshoot hard drive failures using SMART data
https://forums.freenas.org/index.ph...bleshooting-guide-all-versions-of-freenas.17/

Identify your drives by serial number.
Multiple ways to identify your disks for replacement.
https://forums.freenas.org/index.php?resources/identify-your-drives-by-serial-number.64/

Replacing a failed/failing disk
Click-by-click, how to replace a failed or failing disk
https://forums.freenas.org/index.php?resources/replacing-a-failed-failing-disk.75/

 

[Chris Moore]

I've already had to replace one of the drives a year or so ago when it failed. I now have another drive that is failing but has not yet completely failed.

PS. If you have a drive going bad, you should replace it immediately.

 

[danb35]

can someone point me to the process

It's just like the process in the manual (or in my resource), except that you don't offline the failing disk first. No loss of redundancy.

 

[RonRN18]

In my setup, I have three different RAIDZ1 volumes. Each volume is made up of four identical disks. For a case, I have a large case that has 20 hot-swappable drive caddies; set up in five rows of four. The top row is "Volume 1" and all the connections are to four of the motherboard's SATA ports. The next two rows are connected to my first RAID controller (flashed to IT). These are for "Volume 2" and "Volume 3". Both Volumes 2 and 3 are HGST 4TB NAS drives. The last two rows are to a second RAID controller (also flashed to IT). The first volume is where I have my failing drive. Currently, I have no drives in the lower two rows, but I set it up for future expansion. It was such a pain to route the SAS and power cables to the hot swap backplane, I only wanted to do that once when I moved everything into this new behemoth case. Since "Volume 1" was going to 4 SATA ports on the motherboard, I had to use a reverse mini-SAS cable, as the backplane only accepts mini-SAS connections.

In case anyone is curious why I chose RAIDZ1 instead of a more robust/redundant setup, while I don't want to lose any data, the majority of it is ripped Blu-Ray movies that I own. My FreeNAS is primarily used as a Plex Media Server. I have over 600 movie titles. I also have a Sickbeard/SABNZBD setup to function as a huge DVR. I only have about 1.5 TB of personal files and of those 1.5 TB personal files, I have those set to "rclone sync" nightly with Backblaze B2 for offsite backup.

In "Volume 1", all four are WD Red 3TB drives. I've already had one drive completely fail in 2017, which was "ada3". In that situation, I had no choice but to follow the instructions of the official FreeNAS User Guide. This was also before I migrated my system to the case I just described. I now have "ada2" that has twice now given me bad sector warnings. With my new setup, I'm hoping it will be much easier, especially since I don't need to power-down the system for the drive exchange. I'm just debating how to best accomplish it, as "ada2" hasn't completely failed and degraded the system yet.

 

[danb35]

I'm just debating how to best accomplish it,

If you have a spare bay that's active, install a disk there, make sure it's burned in and tested thoroughly, and then start the replacement process through the GUI. Once resilvering is complete, remove the old disk.

 

[Chris Moore]

I'm just debating how to best accomplish it, as "ada2" hasn't completely failed and degraded the system yet.

I thought you didn't want the system degraded? Don't wait for it to fail, you can do a replace on a perfectly healthy drive if you just want to. Pop a replacement drive in one of your unused bays, then you can select it as the replacement and away it goes.

 

[Chris Moore]

make sure it's burned in and tested thoroughly,

Oh, yes, do that first. I always keep spares on hand and I always test them when I get them so they are ready to put in when I need them.