adding disk to zfs mirror after removal of disk

[David Christensen]

On 8/23/20 7:02 PM, tech-lists wrote:

> In the end, I deleted the pool and re-created it ...

Deleting and recreating definitely has its advantages.


> Seems what I should have done is zpool replace ada3 ada3 after fixing the
> failing disk, in order to retain the mirror.

The "diskid..." nodes incorporate drive make, model, and serial number, 
which is appealing because the node name is machine-generated 
(precluding typographic errors), unique, and matches the physical label 
on the drive.  This is probably the best answer if you have a lot of 
disks.  So (see [1]):

# zpool replace wd ada3 diskid/DISK-WD-WCC...


But deleting the pool and starting over gives you the opportunity to 
create and use human-friendly labels.  I have a dozen or so data and 
backup disks, and I use GPT with labels.


It is nice to have multiple good options.  :-)


> Funny you mention this, I'm doing ... cvs too ...  on a dedicated machine.

A networked version control system is very useful for system 
administration, among other things.  I don't know how people get by 
without one.


My CVS repository is in a jail on my SOHO server, along side a jail for 
Samba.  Every time I work on that machine, I lose CVS support.  I have 
contemplated moving CVS to my old laptop for just this reason.  But what 
I really want is to configure another server identically and set up HA 
for both services.


> thanks for your help,

YW.  I'm glad it worked out.  :-)


David


[1] https://docs.oracle.com/cd/E19253-01/819-5461/gbcet/index.html