Retry once and only once if CPI receives AzureConflictError when deleting a managed disk

[bingosummer]

• Please check this box once you have run the unit tests

  pushd src/bosh_azure_cpi
    bundle install
    bundle exec rspec spec/unit/*
  popd
  

The coverage is changed from 2483 / 2724 LOC (91.15%) covered to 2492 / 2733 LOC (91.18%) covered.

Changelog

• When deleting managed disks, Azure may return an AzureConflictError with no retry_after if the disk is used by another operation. This commit is a workaround to handle the error. CPI will retry once and give up if it still fails. After Azure adds retry-after in the response header, the workaround can be removed because the retry in azure_client2 will be triggered by retry_after.

[cfdreddbot]

Hey bingosummer!

Thanks for submitting this pull request! I'm here to inform the recipients of the pull request that you and the commit authors have already signed the CLA.