zpool: fix spare/cache indentation in list -v output

[kevans91]

Motivation and Context

Normal vdevs come out right because of recursion into collect_list_stats with a greater depth, but spare/cache details are emitted at the top-level and need a bit more of a nudge in indentation to avoid blending in with another zpool.

Some details omitted for brevity. Before:

$ zpool list -v -o name,size,alloc,free,health
NAME             SIZE  ALLOC   FREE    HEALTH
storage         43.7T  2.66M  43.7T    ONLINE
  raidz2-0      21.8T  1.30M  21.8T    ONLINE
    [...]
  raidz2-1      21.8T  1.36M  21.8T    ONLINE
    [...]
spare               -      -      -         -
  gpt/disk6     3.64T      -      -     AVAIL
  gpt/disk7     3.64T      -      -     AVAIL
  gpt/disk14    3.64T      -      -     AVAIL
zroot            216G  1.40G   215G    ONLINE
  mirror-0       216G  1.40G   215G    ONLINE
    ada0p3       217G      -      -    ONLINE
    ada1p3       217G      -      -    ONLINE

After:

$ zpool list -v -o name,size,alloc,free,health
NAME             SIZE  ALLOC   FREE    HEALTH
storage         43.7T  2.66M  43.7T    ONLINE
  raidz2-0      21.8T  1.30M  21.8T    ONLINE
    [...]
  raidz2-1      21.8T  1.36M  21.8T    ONLINE
    [...]
  spare             -      -      -         -
    gpt/disk6   3.64T      -      -     AVAIL
    gpt/disk7   3.64T      -      -     AVAIL
    gpt/disk14  3.64T      -      -     AVAIL
zroot            216G  1.40G   215G    ONLINE
  mirror-0       216G  1.40G   215G    ONLINE
    ada0p3       217G      -      -    ONLINE
    ada1p3       217G      -      -    ONLINE

Description

Give print_line() the ability to print indentation and take it into account for the name and use the appropriate depth for both, accounting for the fact that we're missing a level of depth being at the top-level.

How Has This Been Tested?

Ran on FreeBSD 15.1 with spares configured, doesn't really affect non-spare setupss.

Types of Changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Performance enhancement (non-breaking change which improves efficiency)
• Code cleanup (non-breaking change which makes code smaller or more readable)
• Quality assurance (non-breaking change which makes the code more robust against bugs)
• Breaking change (fix or feature that would cause existing functionality to change)
• Library ABI change (libzfs, libzfs_core, libnvpair and libzfsbootenv)
• Documentation (a change to man pages or other documentation)

Checklist

• My code follows the OpenZFS code style requirements.
• I have updated the documentation accordingly.
• I have read the contributing document.
• I have added tests to cover my changes.
• I have run the ZFS Test Suite with this change applied.
• All commit messages are properly formatted and contain Signed-off-by.

[amotin]

What's about log, dedup and special vdevs? They left without indentation. Also spare/cache is not really a vdev layer to have single indentation same as some raidz2. I feel we need some more universal scheme.

[kevans91]

What's about log, dedup and special vdevs? They left without indentation. Also spare/cache is not really a vdev layer to have single indentation same as some raidz2. I feel we need some more universal scheme.

This was purely a drive-by of the obvious ones as I was setting up a new pool, as the current formatting is a little misleading without any vertical spacing between pools. I guess my proposal would be bumping vdevs out by another two and using a vdev label before them?

[behlendorf]

Also spare/cache is not really a vdev layer to have single indentation same as some raidz2.

@amotin I'm probably misunderstanding your comment, but I'd expect log, cache, spare, dedup and special vdevs all to be shown as on the same level as mirror/raidz/draid vdevs. These are all top-level vdevs from a user perspective and should be shown nested under a pool. This is also how the vdev tree is represented in the --json output.

For example:

# zpool list -v 
NAME         SIZE  ALLOC   FREE  CKPOINT  EXPANDSZ   FRAG    CAP  DEDUP    HEALTH  ALTROOT
tank        21.8T   480K  21.8T        -         -     0%     0%  1.00x    ONLINE  -
  mirror-0  7.27T      0  7.27T        -         -     0%  0.00%      -    ONLINE
    U70     7.28T      -      -        -         -      -      -      -    ONLINE
    U71     7.28T      -      -        -         -      -      -      -    ONLINE
  dedup         -      -      -        -         -      -      -      -         -
    U74     7.28T      0  7.27T        -         -     0%  0.00%      -    ONLINE
  special       -      -      -        -         -      -      -      -         -
    U75     7.28T   480K  7.27T        -         -     0%  0.00%      -    ONLINE
  logs          -      -      -        -         -      -      -      -         -
    U76     7.28T      0  7.27T        -         -     0%  0.00%      -    ONLINE
  cache         -      -      -        -         -      -      -      -         -
    U73     7.28T      0  7.28T        -         -     0%  0.00%      -    ONLINE
  spare         -      -      -        -         -      -      -      -         -
    U72     7.28T      -      -        -         -      -      -      -     AVAIL

And we'd want to give zpool status the same treatment.

# zpool status -v
  pool: tank1
 state: ONLINE
config:

	NAME        STATE     READ WRITE CKSUM
	tank1       ONLINE       0     0     0
	  mirror-0  ONLINE       0     0     0
	    U70     ONLINE       0     0     0
	    U71     ONLINE       0     0     0
	  dedup	
	    U74     ONLINE       0     0     0
	  special	
	    U75     ONLINE       0     0     0
	  logs	
	    U76     ONLINE       0     0     0
	  cache
	    U73     ONLINE       0     0     0
	  spares
	    U72     AVAIL   

errors: No known data errors

[amotin]

I'd expect log, cache, spare, dedup and special vdevs all to be shown as on the same level as mirror/raidz/draid vdevs. These are all top-level vdevs from a user perspective and should be shown nested under a pool.

@behlendorf I agree that showing them with the same indentation as a pool is wrong. But neither I particularly like making the class names look as top-level vdevs, since the top-level vdevs in that case are U72-U77.

[kevans91]

I'd expect log, cache, spare, dedup and special vdevs all to be shown as on the same level as mirror/raidz/draid vdevs. These are all top-level vdevs from a user perspective and should be shown nested under a pool.

@behlendorf I agree that showing them with the same indentation as a pool is wrong. But neither I particularly like making the class names look as top-level vdevs, since the top-level vdevs in that case are U72-U77.

Would it help to stylize them differently or something? <cache>, etc.

Open to suggestions- I feel like I've stepped into a small hornet's nest that I'm not entirely qualified to poke at personally, but I'm willing to enact whatever you feel is best.

[behlendorf]

But neither I particularly like making the class names look as top-level vdevs

That's fair, making them class names look like top-level vdevs is misleading as well. I'm not sure I have a great suggestion for how to improve the output. I'm open to suggestions, but until we can a concrete suggestion we're happy with I'd suggest we close this out.