zedmanager: do not let a domain-held stale volume ref block a purge

[andrewd-zededa]

Description

doInstall keeps a VolumeRefStatus that the current AppInstanceConfig
no longer references around, unremoved, for as long as the running domain
still uses it — so its release can be ordered after the domain is torn
down rather than yanking storage out from under a live app. That part is
correct.

But doInstall's aggregate error computation still folded that stale
entry's error into the app's own blocking error, so doInstall never
returned done=true while the stale volume was in an error state. That
deadlocks the purge: doUpdate bails out on !done before ever reaching
the code that tears down the domain — which is the only thing that lets
the stale volume actually go away. A volume that failed once (independent
of any purge) and then had a purge move the app to a new generation would
wedge that purge forever, with no automatic path out.

Fix: skip a stale, domain-held VolumeRefStatus's error (and state) when
computing doInstall's aggregate result during the DownloadAndVerify
phase — only errors on volume refs still present in the current config
should be able to block progress.

Found while debugging two apps on an eve-k dev cluster stuck indefinitely
in DownloadAndVerify after an unrelated volume-creation failure (a
separate CDI-upload-completeness bug) left one volume permanently errored
right as a purge tried to move the app to a new generation.

PR dependencies

None

How to test and validate this PR

Covered by an automated regression test:

make TESTPKGS=./pkg/pillar/cmd/zedmanager/... TESTRUN=TestDoInstall test

Covered by an automated regression test:

make TESTPKGS=./pkg/pillar/cmd/zedmanager/... TESTRUN=TestDoInstall test

TestDoInstallErroredStaleVolumeRefHeldByDomainDoesNotBlock reproduces the
exact shape: a domain still attached to an errored old-generation volume,
plus a new-generation volume already LOADED and waiting on VerifyOnly
to clear. Before the fix, doInstall returns done=false forever; after
it, done=true and the stale volume's error is not folded into the app's.

Changelog notes

Fixed a bug where an application purge on eve-k could hang indefinitely
if the app's currently-active volume was in a permanent error state when
the purge was issued — the purge would never tear down the old app
generation or start the new one. No other user-facing changes.

PR Backports

• 17.0-stable: To be backported.
• 16.0-stable: No, as the feature is not available there.
• 14.5-stable: No, as the feature is not available there.
• 13.4-stable: No, as the feature is not available there.

Checklist

• I've provided a proper description
• I've added the proper documentation
• I've tested my PR on amd64 device
• I've tested my PR on arm64 device
• I've written the test verification instructions
• I've set the proper labels to this PR

And the last but not least:

• I've checked the boxes above, or I've provided a good reason why I didn't
  check them.

Please, check the boxes above after submitting the PR in interactive mode.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 26.04%. Comparing base (7f50511) to head (7ef13c3).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #6442      +/-   ##
==========================================
+ Coverage   25.67%   26.04%   +0.37%     
==========================================
  Files         509      519      +10     
  Lines       93304    95023    +1719     
==========================================
+ Hits        23955    24748     +793     
- Misses      67501    68248     +747     
- Partials     1848     2027     +179     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.