Gravity 7.0.12 LTS / 6.1.30 LTS Now Available

7.0.12 LTS (July 10th, 2020)

Improvements

  • Update gravity status to display unhealthy critical system pods (#1753, planet#691, satellite#236).
  • Add --since flag to gravity report command to allow filtering collected logs by time (#1806).
  • Add more system information collectors to gravity report (#1806).
  • Set priority classes on critical system pods (#1814, logging#66, monitoring#181).
  • Update etcd disk check to produce a warning on soft limits (50ms latency, 50 IOPS) and a critical failure on hard limits (150ms latency, 10 IOPS) (#1847).

Bugfixes

  • Make sure upgrade is resumed on the correct node to avoid issues with plan inconsistency (#1794).
  • Make sure upgrade can’t be launched if the previous upgrade wasn’t completed or fully rolled back (#1803).
  • Fix an issue with upgrade failing to adjust volume permissions if only uid is specified (#1803).
  • Fix an issue with some tabs not working when accessed through wizard installer user interface (#1815).
  • Fix an issue with installation failing on Vultr (#1815).
  • Fix an issue with installed Helm releases not displayed on the dashboard (#1815).
  • Fix an issue with signup/reset links not honoring configured public address (#1815).
  • Fix an issue with tsh sometimes producing access denied errors (#1836).

6.1.30 LTS (July 10th, 2020)

Improvements

  • Disk space check will now trigger a warning on soft limit (80%) and degrade the node on hard limit (90%) (#1690, planet#678, #228).
  • Add backoff after new cluster image upload to make sure cluster is healthy (#1709).
  • Display a more detailed reason for degraded status in gravity status (#1723).
  • Update gravity status to display unhealthy critical system pods (#1752, planet#690, satellite#235).
  • Upgrade Grafana to v6.7.4 (#1754, monitoring#177).
  • Add --since flag to gravity report command to allow filtering collected logs by time (#1800).
  • Add more system information collectors to gravity report (#1800).
  • Update ClusterConfiguration resource to support service CIDR changes (#1839).

Bugfixes

  • Fix an issue with monitoring app upgrade sometimes failing with “request too large” (#1697, monitoring#174).
  • Fix an issue with kube-apiserver sometimes entering failed state during failover (#1742, planet#685).
  • Fix an issue with calculating port differences when upgrading to a new version (#1745).
  • Fix an issue with RPC agent credentials not being rotated during upgrade (#1747).
  • Make sure upgrade is resumed on the correct node to avoid issues with plan inconsistency (#1793).
  • Make sure upgrade can’t be launched if the previous upgrade wasn’t completed or fully rolled back (#1802).
  • Fix an issue with upgrade failing to adjust volume permissions if only uid is specified (#1802).