Fix deadlock and DELETE/UPDATE unindexed datasets

[dbutenhof]

PBENCH-1328

A bug revealed another bug:

1. POST/DELETE /datasets/{id} fails when the dataset isn't indexed; with
   the ability to disable server indexing entirely and to remove dataset indexed
   data, this is undesirable. As long as the dataset doesn't have a WORKING ops
   status, we should be able to update or delete.
2. When the operation fails, the status is already set to WORKING, and this
   was not corrected on exit, leaving the dataset locked perpetually.

Fixed in two phases: first to correct the handling of errors so that we'll
always unlock on failure; second to allow deletion of unindexed datasets. (For
PR review, these phases are available in separate commits.)

[dbutenhof]

I was hoping to do some testing on a runlocal but for some reason I can't quite determine, the Keycloak container is working as expected, and the setup times out. If the CI runs, that'll be a useful diagnostic.

(And I probably should have posted as a Draft, because I realize that my second commit wiped out explicit testing of the WORKING state fix, which will now require some mocking to set up; but I'm also not sure I want to be spending more time on this. We'll see if I get bored over the weekend.)

[dbutenhof]

Ah: so I think I found our problem:

$ openssl x509 -enddate -noout -in server/pbenchinacan/etc/pki/tls/certs/pbench_CA.crt 
notAfter=Apr 25 21:37:54 2024 GMT

@webbnh Do you remember exactly what you did originally to generate the CA? (It should probably be recorded in a README.md in the pbenchinacan directory.)

Nevermind: I figured it out. I'm going to post a new commit with a new CA, but I'm also going to need to fix some stuff so I'm changing to draft.

[webbnh]

Looks good. I have one pointed question, and a few suggestions which probably aren't worth pursuing...unless you want to. 🙃

[webbnh]

On the previous pass, I neglected to review all of the commits.... I found a few typo-things that you might want to fix.

[webbnh]

👍