[Ingest Manager] Fix failing installation on windows 7

[michalpristas]

What does this PR do?

Issue described here: #24327
there was a race between enrollment process and restarting service, FS playing part as well.
The thing was that when agent was restarted only on windows 7 it loaded standalone ID, even though it was already replaced by enrollment process.
Then when agent retrieved hosts from fleet it even overwrote updated ID with stale one.

This fix adds a lock which prevents simultaneous write in between these two processes and a forced Reload in case of fleet managed agent later in the cycle.
Another thing is FSync after file rotation which was missing for windows.

These seems to fix the issue, tested on win 7 VM on cloud

Why is it important?

Fixes #24327

Checklist

• My code follows the style guidelines of this project
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• I have made corresponding change to the default configuration files
• I have added tests that prove my fix is effective or that my feature works
• I have added an entry in CHANGELOG.next.asciidoc or CHANGELOG-developer.next.asciidoc.

[elasticmachine]

Pinging @elastic/ingest-management (Team:Ingest Management)

[elasticmachine]

Pinging @elastic/agent (Team:Agent)

[elasticmachine]

💚 Build Succeeded

the below badges are clickable and redirect to their specific view in the CI or DOCS

Expand to view the summary

Build stats

• Build Cause: Pull request #24387 updated

• Start Time: 2021-03-07T20:55:21.137+0000

• Duration: 51 min 28 sec

• Commit: ff4bee8

Test stats 🧪

Test	Results
Failed	0
Passed	45789
Skipped	4916
Total	50705

Trends 🧪

💚 Flaky test report

Tests succeeded.

Expand to view the summary

Test stats 🧪

Test	Results
Failed	0
Passed	45789
Skipped	4916
Total	50705

[blakerouse]

I think an infinite loop is present.

[blakerouse]

What if this always fails? How does this not loop forever?

[michalpristas]

yeah copy paste i rewrote that part

[blakerouse]

Retry logic looks correct now.

[dikshachauhan-qasource]

Hi @EricDavisX

Today we have validated the above fixes on upgraded Kibana from 7.10.2 to 7.12 and found the issue as fixed.

Installed agent with only system integration in Agent policy:

Observations:

• Agent is in healthy state.
• Activity logs present
• logs on discover tab are visible.
• Metricbeat and filebeat are running.

Build details:

BUILD 39134
COMMIT 08417cbd6c15e4c866651a7dcdfeded58845206d
Artifact link: https://staging.elastic.co/7.12.0-96914cb5/summary-7.12.0.html

Thanks
QAS