Smartctl notification of "unreadable pending sector"

[crudolphy]

I have a Dell T320 server. I have 8 hard drives. One is dedicated to Proxmox VE version 8.2.2. I have 3 drives in one zfs pool using raidz1 the pool name is "tank-1". I get an email notification everyday saying that one of the drives "Device: /dev/sdb [SAT], 1 Currently unreadable (pending) sectors." It is a Hitachi HDS721010CLA332 1TB drive. The other 2 drives in the pool are identical mfg, model, and size. There is also an entry in the proxmox syslog of "unreadable pending sector" for /dev/sdb.

I ssh'd into the box and ran the command "smartctl -l selftest /dev/sdb". it returned no errors. i then ran the command "zpool scrub tank-1" (tank-1 is the zpool name). it returned no errors.

When in the Proxmox admin gui I go to the node, then zfs and it lists both my pools. When I double click on the pool tank-1, it shows the pool as online and healthy (green checkmark) then it shows that it is configured as raidz1-0 as online and healthy, then it shows the three physical disks that make up the pool and all show online and healthy. The columns with read, write, and checksum all show zeros and the column for messages is blank (no messages).

I found (2) Hitachi with the same model # on ebay for $24.00 USD so I bought them.

Question #1. Do I have a drive failing?
Question #2. If the answer to #1 is NO then how do I clear up the logs and how do I stop the email message but still continue to get newer ones?
Question #3. If the answer to #1 is YES then can you point me to a tutorial or expicit instructions to replace the drive.

I have searched youtube for a video and can't find one that explains it sufficiently (there are many for true nas [none from the command line] not so much for proxmox). I looked at the proxmox documentation and the only thing I could find was this command "zpool replace -f <pool> <old-device> <new-device>". Is it that simple? What are they referring to when they say <old-device> <new-device>?

I sure could use some answers/pointers/tutorial/explanations and would really appreciate anyone's help.

Regards

Chuck

 

[Kingneutron]

Q1: Yes.

Q3: https://www.dlford.io/linux-zfs-raid-disk-replacement-procedure/
^ You can skip steps 3 and 4, zfs creates partitions 1 and 9 when you attach a full disk

If the system gave the replacement disk a new short drive letter e.g. sdz , you would do:

zpool replace tank-1 /dev/sdb /dev/sdz

Where sdb is the failed/removed device. ' zpool status -v ' may give you a long number instead of a device name to replace, use that in the replace command if so.

 

[Dunuin]

I looked at the proxmox documentation and the only thing I could find was this command "zpool replace -f <pool> <old-device> <new-device>". Is it that simple? What are they referring to when they say <old-device> <new-device>?

Yes, it is when PVE isn't installed on that pool. "<old-device>" is the disk you want to replace, "<new-device>" is your disk you want the existing one to be replaced with.

 

[crudolphy]

Thanks guys! I really appreciate your responses. Not going to try tonight but will tomorrow. Let you know how it comes out. Regards
Chuck