Addressed issues¶

The MCP 2019.2.22 update contains fixes for the following MCP components:

• DriveTrain

  • [37043] Fixed the issue with the incorrect order of Galera and MySQL packages installation during the Galera cluster update and during the Upgrade OS stage.

  • [37013] Fixed the issue causing the NGINX failure due to the configuration file containing the dot symbols in the location section.

  • [36930] Fixed the high CPU loads issue in environments with NFS storage enabled by making the image_cache_manager_interval parameter configurable.

    Note

    Within Mirantis Cloud Platform, the parameter defaults to 0, but we recommend following the upstream default value of 2400. For details about the image_cache_manager_interval parameter, see OpenStack documentation: Configuration options.

  • [35604] Fixed the issue with QEMU overconsuming system resources by adding the following options:

    • qemu_max_processes

      Sets the maximum number of processes that a QEMU user can run. This option overrides the default value set by the host OS.

    • qemu_max_files

      Sets the maximum number of files that a QEMU user can open.

• OpenStack:

  • [37023][Pike, Queens] Updated dnsmasq to the most recent Ubuntu Bionic release to eliminate USN-4976-2: Dnsmasq vulnerability.
  • [37019][Pike, Queens] Fixed the Neutron traffic overloading issue by fetching a new upstream version of the Galera library.
  • [37017][Pike] Fixed the instance_fault table overloading issue by adding InstanceInvalidState to the table exceptions.
  • [37014][Pike, Queens] Fixed the issue causing ldappool to retry an incorrect password until reaching the retry limit.
  • [37012][Pike, Queens] Updated the MySQL wsrep package to version 5.7.39 to include the upstream security fixes.