zpool: Add slot power control, print power status

[tonyhutter]

Motivation and Context

Use zpool commands to control enclosure and NVMe slot power. Sometimes it's useful to power off the slot to a misbehaving drive.

Description

Add zpool flags to control the slot power to drives. This assumes your SAS or NVMe enclosure supports slot power control via sysfs.

The new -0|-1 flags are added to zpool offline|online|clear:

    zpool offline -0 <pool> <device>    Turn off device slot power
    zpool online -1 <pool> <device>     Turn on device slot power
    zpool clear -1 <pool> [device]      Turn on device slot power

If the ZPOOL_AUTO_POWER_ON_SLOT env var is set, then the -1 is automatically implied for zpool online and zpool clear and does not need to be passed.

Also add a zpool status -1 option to print the slot power status as a new column.

Note: This also adds a FOR_EACH_VDEV()/FOR_EACH_LEAF_VDEV() macro that makes iterating over vdevs a lot easier.

How Has This Been Tested?

Tested locally on NVMe and SAS enclosures.

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)
• Breaking change (fix or feature that would cause existing functionality to change)
• Library ABI change (libzfs, libzfs_core, libnvpair, libuutil 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.

[allanjude]

I think the command line interface could be clearer, like --power or something, since on/off is implied by online vs offline (or maybe --power=[on|off] for clear?), rather than just -1 and -0

[AllKind]

I'd second that.
Unfortunately zfs does not use --long-options.

	online [-e] <pool> <device> ...
	offline [-f] [-t] <pool> <device> ...
	clear [-nF] <pool> [device]
	status [-c [script1,script2,...]] [-igLpPstvxD]  [-T d|u] [pool] ... 

Obvious -pP is taken by status, -e (engergy) by online.
Maybe -w (for watt - as the closest?).

[robn]

Unfortunately zfs does not use --long-options.

A few commands do. Not hard to add it in to those that don't.

[AllKind]

Ok, should have said zpool clear, offline, online, status...

[tonyhutter]

I think the command line interface could be clearer, like --power or something

What would the short name for --power be?

Maybe -w (for watt - as the closest?).

If we want to later add power control to zpool remove (which is possible) then -w is not going to work, since it is already taken for zpool remove. -w is also used by zpool iostat.

[robn]

What would the short name for --power be?

Does it need one? Depends how niche this feature is, I suppose.

[tonyhutter]

What would the short name for --power be?

Does it need one? Depends how niche this feature is, I suppose.

I suppose not. We have short names for all the other long args, but there's nothing set in stone that we have to have them.

[robn]

There's one we don't: zpool import --rewind-to-checkpoint. Just in case you didn't want to be the one setting the precedent 😉

[tonyhutter]

@robn thanks, that settles it then - I'll add it as --power.

[tonyhutter]

All the changes requested have been pushed in. I'm still looking into some build warnings on Centos 7 and FreeBSD.

[tonyhutter]

I think this is ready to go. The checkstyle failure does not appear for me when I run make checkstyle on F39, so I'm not sure what's going on with that.

[tonyhutter]

My latest push uses the ABI files from the checkabi artifacts, which resolves the errors.