Document resetting Rock-on Root (#434)

[Hooverdan96]

Fixes #434 .

This pull request's proposal (superseeding PR #502).

add the Rock-on root reset process to the section that deals with uninstalling/installing Rock-ons. Also, some minor cleanups in the same file.

Checklist

• With the proposed changes no Sphinx errors or warnings are generated.
• I have added my name to the AUTHORS file, if required (descending alphabetical order).

Note: Linkcheck will continue to fail until a currently open PR has been merged.

[phillxnet]

@Hooverdan96 Re:

Note: Linkcheck will continue to fail until a currently open PR has been merged.

This also requires that this branch be re-based. But not required as we know the failure now. And I have now merged that PR. So we are hopefully getting ahead of the little thorn.

About to review this now so no need to re-base by the way.

[phillxnet]

@Hooverdan96 Thanks for yet-another great addition.

I'm a bit concerned about folks jumping to this brute-force solution: but there are warnings in there. But this is rather the nuclear option in my opinion. But as you indicated here, it can be necessary on occasions. So it can only be good to have it documented as you've presented. Thanks.

See what you think of my suggested changes, almost all trivial actually; but hopefully also nice-to-have.

My main indicator that this this procedure is necessary has been log entries to the effect that a subvol/snapshot was was missing pertaining to a docker container. What we likely need in rockstor-core here is more use of:

docker system prune -a --volumes

And maybe that could be added to this doc as a pre-nuclear (wipe / recreate) option. And only if that, once run, followed by a Pool refresh to update subvols fails to work do we then suggest this rather radical option of switching/refreshing the Rock-ons root share.

[Hooverdan96]

Thanks for the suggestions, I've accepted them. I will incorporate your additional comments from above and bring these back (and address the current Sphinx failure that now seems to have popped up).

[phillxnet]

@Hooverdan96 Thanks. Let me know when I should re-test and publish:
Re the Sphinx error:

Reset The Rock-ons Root
------------------ [docutils]
/home/runner/work/rockstor-doc/rockstor-doc/interface/overview.rst:503: WARNING: Title underline too short.

Yes, that one was down to me I'm afraid. I failed to also modify the number of chars indicating a sub-header to match the sub-header changes I made. My apologies for the extra work there.

[Hooverdan96]

@phillxnet, I did end up rebasing the branch, and rearranged the nuclear option to be after a section on pruning, and also updated the logs symptom with your comment about missing subvols.

[Hooverdan96]

The linkcheck failures this time around where related to smartmontools links, which timed out (first time for me). i assume those are temporary at this point:

(    howtos/smart: line    6) timeout   https://www.smartmontools.org/ - HTTPSConnectionPool(host='www.smartmontools.org', port=443): Read timed out. (read timeout=20)
(    howtos/smart: line  191) timeout   https://www.smartmontools.org/browser/trunk/smartmontools/smartctl.8.in - HTTPSConnectionPool(host='www.smartmontools.org', port=443): Read timed out. (read timeout=20)
(    howtos/smart: line   94) timeout   https://www.smartmontools.org/browser/trunk/smartmontools/smartd.conf.5.in - HTTPSConnectionPool(host='www.smartmontools.org', port=443): Read timed out. (read timeout=20)
(    howtos/smart: line  193) timeout   https://www.smartmontools.org/wiki/Supported_RAID-Controllers - HTTPSConnectionPool(host='www.smartmontools.org', port=443): Read timed out. (read timeout=20)

[phillxnet]

@Hooverdan96 Thanks for including that extra command, may end-up saving some folks from having to do the more extensive reset procedure.

I'll merge and get this published. As always much appreciated. It's great to have this procedure finally documented. In time we can automate better a Rock-on reset, or have Web-UI intergration for force uninstall and the like; but bit-by-bit.

[phillxnet]

Re the smartmontools linkcheck hick-up, a couple of builds locally did not show this failure. So as you surmised I think we caught them on a bad day. Or they don't like GitHub actions accessing their site.

[phillxnet]

PR product PRODUCTION published.