20.10.5

(2021-05-17)

Components

Component

Version

Mirantis Container Runtime

20.10.5

containerd

1.4.4

runc

1.0.0-rc92

Packaging

  • Added a technical preview for cri-docker (previously known as dockershim).

  • Fixed an issue wherein killing containerd caused systemd to stop docker.

Client

  • Reverted docker/cli#2960 to stop hanging in docker start --attach and removed spurious Unsupported signal: <nil>. Discarding messages (docker/cli#2987).

Builder

  • Fixed an issue wherein the classic builder was silently ignoring unsupported Dockerfile options. It now prompts the user to enable BuildKit instead (moby/moby#42197).

Logging

  • Fixed sporadic and unexpected EOF errors in json-file handling (moby/moby#42174).

Networking

  • Fixed a regression that caused IPv6 addresses not to be bound by default when mapping ports (moby/moby#42205).

  • Fixed implicit IPv6 port-mappings not included in the API response. Prior to version 20.10.0, while published ports were accessible through both IPv4 and IPv6 by default, the API only included information about the IPv4 (0.0.0.0) mapping (moby/moby#42205).

  • Fixed a regression wherein the docker-proxy did not terminate in all cases (moby/moby#42205).

  • Fixed an issue wherein iptables forwarding rules were not cleaned up when containers were removed (moby/moby#42205).

  • Fixed an issue wherein swarm service VIPs timed out from Windows containers, resulting in the following error in dockerd event logs:

    Failed to add ILB policy for service service [] failed during
    hnsCallRawResponse: hnsCall failed in Win32: The specified port already
    exists.
    

    The Windows Operating System update KB4577668 introduced the issue on October 13, 2020, affecting all versions of MCR (FIELD-3310).

Known issues

  • Centos and RHEL 7.x kernels may experience memory exhaustion due to slab cache usage. Because this is a kernel feature, the issue cannot be resolved with MCR code.

    Workarounds:

    • (Recommended) Update to a kernel that supports setting cgroups.memory=nokmem and apply the change (for customers using Centos or RHEL 7.7 and above).

    • Increase the memory available to the machine for slab usage.

    (FIELD-3466)