Skip to content

v1.1.15

Compare
Choose a tag to compare

1.1.15 (July 13, 2022)

BUG FIXES:

  • api: Fixed a bug where Consul token was not respected for job revert API [GH-13065]
  • cli: Fixed a bug in the names of the node drain and node status sub-commands [GH-13656]
  • client: Fixed a bug where max_kill_timeout client config was ignored [GH-13626]
  • cni: Fixed a bug where loopback address was not set for all drivers [GH-13428]
  • core: Fixed a bug where an evicted batch job would not be rescheduled [GH-13205]
  • core: Fixed a bug where reserved ports on multiple node networks would be treated as a collision. client.reserved.reserved_ports is now merged into each host_network's reserved ports instead of being treated as a collision. [GH-13651]
  • core: Fixed a bug where the plan applier could deadlock if leader's state lagged behind plan's creation index for more than 5 seconds. [GH-13407]
  • csi: Fixed a regression where a timeout was introduced that prevented some plugins from running by marking them as unhealthy after 30s by introducing a configurable health_timeout field [GH-13340]
  • csi: Fixed a scheduler bug where failed feasibility checks would return early and prevent processing additional nodes [GH-13274]
  • lifecycle: fixed a bug where sidecar tasks were not being stopped last [GH-13055]
  • ui: Allow running jobs from a namespace-limited token [GH-13659]
  • ui: Fixed a bug that prevented the UI task exec functionality to work from behind a reverse proxy. [GH-12925]
  • volumes: Fixed a bug where additions, updates, or removals of host volumes or CSI volumes were not treated as destructive updates [GH-13008]