v2.2.53

This release includes all changes from v2.2.52 plus the ones listed below.

Bug fixes:

• Start exthandler with the same python interpreter (#2007)
• Verify that the extension status is an array (#2010)
• Remove enum _UpdateType and retry fetching goal state (#2018)
• use dd for ext4 as well as xfs (#2042)
• Fix path for error.json (#2044)
• Switch to run command changes, + provisioning changes that need to be reverted. (#2050)
• Fix timestamp for goal state archive (#2051)
• Case insensitive parsing or Plugins and PluginSettings (#2054)
• Revert "Fixed delays for HTTP retries rather than exponential delays (#1967)" (#2065)
• Fixed bug causing "MAC verified OK" message (#2069)
• Revert unicode fix manually (#1937) (#2070)

New features and improvements:

• Recreate handler environment file on service startup (#1960)
• Add log collection tool and thread (#1987)
• Thread interface (#1990)
• Verify that the CPU and Memory cgroups for the agent are properly initialized; disabled cgroups if they are not active. (#2015)
• SUSE config: use Btrfs LZO compression for ResourceDisk (#2055)

Telemetry and logging improvements:

• Extension telemetry pipeline (#1918)
• Reformatted the heartbeat event (#2009)
• Add LIS version to OSInfo.message (#2011)
• One thread for telemetry (#2019)
• Limit description character length sent for health report (#2020)
• Remove Serial Console Logging (#2028)
• Echo log to /dev/console during provisioning (#2043)
• Adding telemetry for logrotate (#2045)
• Report placeholder extension status as an array (#2068)

Documentation updates:

• Fix broken link in readme (#2014)
• Add log collector flags to README (#2029)

Testing and code quality improvements:

• Travis CI improvements (#1989, #1992)
• Test improvements (#1969, #1994, #1988, #2026)
• Enabled pylint (#1975, #1976, #1979, #1980, #1981, #1985)
• Code clean-up, fixes for pylint warnings (#1983, #1986, #1991, #1993, #1995, #1999, #2000, #2001, #2002, #2047, #2049, #2053)

v2.2.51 - this release was not completed in all regions

This release is a re-release of v2.2.49.2. There are no new changes in this as compared to the last known good v2.2.49.2

v2.2.52 - this release was not completed in all regions

Major bug fixes -

• Do not retrieve users in each goal state (#1935)
• Fix check for systemd-run failure when invoking extensions (#1943)
• Fix telemetry unicode errors (#1937)
• Uninstall unregistered extensions (#1970)
• Use run_command to execute iptables (#1944)
• Use run_command for ip route (#1958)
• Fix handling of gen2 disks with udev rules (#1954)

New features/improvements -

• Add API for uploading logs via host plugin (#1902)
• Fixed delays for HTTP retries rather than exponential delays (#1967)

Documentation changes -

• python3 must be used in Ubuntu (#1932)

Other fixes -

• Resolve undefined variable (#1950)
• Convert owner uid to string (#1949)
• Fix Travis special checks for distro and remove useless cgroup tests (#1959)
• Use tmp_dir instead of data_dir (#1968)

v2.2.49.2

Improvements in cgroups:

• Do not use --unit with systemd-cgls (#1910)
• Report processes that do not belong to the agent's cgroup (#1908)
• Use controller mount point for extension cgroup path (#1899)
• Improvements in setup of cgroups (#1896)
• Remove ExtensionsMetricsData and per-process Memory data (#1884)
• Fix return value of start_extension_command (#1927)

Testing and dev-infra improvements:

• Remove import * (#1900)
• Fix flaky ExtensionCleanupTest class (#1898)
• Fix codecov badge (#1883)
• Changed codecov to run on py3.8 (#1875)

Documentation improvements:

• Update documentation on /dev/random (#1909)
• Mount options are in mount(8) (#1893)

Logging and telemetry improvements:

• Remove ssh host key thumbprint in report ready (#1913)
• Emit AutoUpdate value at service start only (#1907)
• Add logging for version mismatch (#1895)
• Send telemetry event if libdir changes (#1897)
• Add log collector utility (#1847)
• Move AutoUpdate reporting to HeartBeat event (#1919)

Extension workflow improvements:

• Removing infinite download of extension manifest without a new GS (#1874)
• Fix wrongful dir deletion (#1873)
• Fix the cleanup-outdated-handlers to only delete handlers that are not present in the GS (#1889)

Other:

• Expose periods of environment thread in waagent.conf (#1891)
• Added user @kevinclark19a as Contributor. (#1906)

v2.2.48.1

Refactoring GoalState class out of Protocol, making Protocol thread-safe, removing stale dependencies of Protocol and removing the dependency on the file system to read the Protocol info:

• Fetch goal state when creating HostPluginProtocol (#1799)
• Separate goal state from the protocol class (#1777)
• Make protocol util a singleton per thread (#1743, #1756)
• Fetch goal state before sending telemetry (#1751)
• Remove file dependency (#1754)
• Others (#1758, #1767, #1744, #1749, #1816, #1820)

Adding and improving log messages and adding thread names to each logline:

• New logs for goal state fetch (#1797) and refresh (#1794).
• Thread name added to logs (#1778)
• Populate telemetry events at creation time (#1791)
• Periodic HeartBeat to be logged to the file (#1755)
• Add unit test to verify call stacks on telemetry events (#1828)
• Others (#1841, #1842, #1846)

Improving agent status reporting:

• Handling errors while reading extension status files (Limiting Size and Transient issues)(#1761)

Provisioning Agent changes:

• Enable SWAP on Resource Disk as Application Certification Support suggested (#1762)
• Update 'Provisioning' options in default configs ( #1853)

Other fixes:

• Drop Metadata Server Support (#1806, #1839, #1840 )
• Improve documentation of ResourceDisk.EnableSwapEncryption (#1782)
• Removed is_snappy function (#1774)
• Handle exceptions in monitor thread (#1770)
• Fix timestamp for periodic operations in the monitor thread (#1879)
• Fix permissions on the Ubuntu systemd service file (#1814)
• Update hostname setting for SUSE distros (#1832)
• Python 3.8 improvements + support for Ubuntu 20.04 (#1860, #1865, #1738)
• Testing and dev-infra improvements [#1771, #1768, #1800, #1826, #1827, #1833]
• Others (#1854, #1858)

v2.2.46

[#1741] Do not update goal state when refreshing the host plugin
[#1731] Fix upgrade sequence when update command fails
[#1725] Initialize CPU usage
[#1716, #1737] Added UTC logging and correcting the format
[#1651, #1729] Start sending PerformanceCounter metrics and additional memory information for Cgroups

v2.2.45

[#1654] Add support for Gen2 VM resource disks
[#1677] Use alternate systemd detection
[#1638] Fix /proc/net/route requirement that causes errors on FreeBSD
[#1633] Add cloud-init auto-detect to prevent multiple provisioning mechanisms from relying on configuration for coordination
[#1688] Disable cgroups when daemon is setup incorrectly
[#1686] Remove upgrade extension loop for the same goal state
[#1696] Add container id for extension telemetry events
[#1695] Be more exact when detecting IMDS service health
[#1663] Changing add_event to start sending missing fields

Bug fixes and minor improvements: [#1679, #1680, #1681, #1683, #1685, #1699, #1698, #1701]

v2.2.44

[#1642] Remove outdated extension ZIP packages
[#1623] Improved error handling when starting extensions using systemd
[#1619] Reduce provisioning time of some custom images
[#1617] Improve the handling of extension download errors
[#1634, #1640] New API for extension authors to handle errors during extension update
[#1606] Fix handling of errors in calls to openssl
[#1599] Improve logic to determine current distro
[#1567, #1614] Reduce verbosity of several logging statements

Bug fixes: [#1506 #1563 #1568 #1570 #1572 #1573 #1576 #1580 #1583 #1595 #1600 #1606 #1607 #1616 #1625 #1628 #1630 #1641 #1644 #1655 #1664 #1666]

v2.2.42

The following release addresses an issue in how the Azure Linux Agent processes goalstates when the extension processing is on hold (either extension processing is disabled in the conf file or the 'on_hold' property is marked as true in the artifact blob).

[#1602] Poll for artifact blob

v2.2.41

This release contains the following changes:

• Rewriting the mechanism to start the extension using systemd-run for systems using systemd for managing
  [#1537, #1536]
• Refactoring of resource monitoring framework using cgroup for both systemd and non-systemd approaches [#1530, #1534]
• Telemetry pipeline for resource monitoring data [#1538]