I just acquired some new hard drives for my ArchiveVolume pool. I always burn new drives in, its a simple process but I thought id go over it with my new site to get some content flowing!
First its best to start with a short smart test, just to confirm the drive is the right drive, and that the drive is ready to go.
I’m on FreeBSD using
gmultipath for my drives (attached over ses).
smartctl can’t use
/dev/multipath/disk<id> so ill have to invoke the
gmultipath list command to discover the disk name.
# gmultipath list Geom name: disk15 Type: AUTOMATIC Mode: Active/Passive UUID: 9f6b9313-f39b-11ea-9ad7-001c23d66829 State: OPTIMAL Providers: 1. Name: multipath/disk15 Mediasize: 2000398933504 (1.8T) Sectorsize: 512 Mode: r1w1e0 State: OPTIMAL Consumers: 1. Name: da58 Mediasize: 2000398934016 (1.8T) Sectorsize: 512 Mode: r2w2e1 State: ACTIVE 2. Name: da59 Mediasize: 2000398934016 (1.8T) Sectorsize: 512 Mode: r2w2e1 State: PASSIVE
Now that i know what names the consumers go by i can run my tests.
# smartctl -t short /dev/da59
If i want to view the contents of the test i can run the following.
# smartctl -a /dev/da59
Remember you need to replace /dev/da59 with the disk you’re trying to access. If you’re not sure you can check the hard drive serial numbers with a tool like
geom disk list.
Now onto the next steps, id recommend running a
conveyance and a
long test next, remember these tests take time to continue, a %left and estimated total time can be seen using
smartctl -a <disk>.
Next is the fun part.
badblocks this program will flip every bit on your disk and its one of my favorites, make sure to start up a
tmux session once you’re ready for this step so an accidental trip over your networking cable won’t cancel your
badblocks command, this can take days!
##Run this in a tmux session! ##This will nuke all data on the drive! Be very sure about running badblocks. # badblocks -ws /dev/multipath/disk15 ##If your drive is newer, the process can be much faster if you specify the right block size # badblocks -b 4096 -ws <disk>
Once that’s done, then that’s it! If your drives survived, then they’re probably going to do so for a while longer. Remember to regularly run SMART tests and keep an eye on drive health!