[pve-devel] [PATCH docs v2 1/6] ceph: add anchors for use in troubleshooting section

[Max Carrara]

On Wed Feb 5, 2025 at 11:08 AM CET, Alexander Zeidler wrote:
> Signed-off-by: Alexander Zeidler <a.zeidler at proxmox.com>
> ---
> v2:
> * add two missing anchors to be usable via xref
>

To keep things short:

- Docs now build again (tested with both `make update` and `make deb`).

- Installed the .deb packages on my development VM in order to read the
  docs where they'll be "deployed" - all works fine.

- Gave anchors and hrefs a smoke test by testing 'em out randomly;
  seems like everything's working .
  (tested on multi-page docs, single-page docs, PDF version)

- Also checked if the `id` attribute is set for the smallest headings,
  and it sure is! You can't link to those sections by default (which is
  an AsciiDoc and/or configuration thing, I'm guessing), *but* we should
  still be able to refer to them throughout the docs, if necessary
  (untested).

  For example, the `id` for "Relevant Logs on Affected Node" is
  set correctly as intended: `chapter-pveceph.html#pve_ceph_ts_logs`

- As mentioned before, I find the writing style to be quite nice; I
  especially like how things are broken down into smaller steps and
  paragraphs. No fancy idioms or figures of speech; the writing is
  strictly technical and instructive.

- The instructions themselves also seem fine; I have used similar steps
  the last time I was messing around with my Ceph cluster (was a while
  ago though). I especially like the update on the "Destroy OSDs"
  section; I personally wouldn't have thought of checking that the OSDs
  should have their Used % below the nearfull_ratio before throwing an
  OSD out.

  Full disclosure: The only thing I haven't tried out is removing a node
  that is running Ceph from a cluster. *But* because I wiped a Ceph
  installation from a node before (before re-creating it again) I can
  tell that the steps there are sensible (and *much* safer than what I
  did back then, woops). The only thing I hadn't done was removing the
  node from the CRUSH hierarchy, but I guess in my case Ceph had just
  figured that out itself :P

All in all, unless I missed something or if there are any objections, I
think this can be merged.

Consider:

Reviewed-by: Max Carrara <m.carrara at proxmox.com>
Tested-by: Max Carrara <m.carrara at proxmox.com>