fix: accept all docker memory units in resource values

Memory values with uppercase or other docker-supported units such as `1G`, `512M` or `1GiB` are now passed to docker unchanged instead of having an extra `m` appended, and the k3s scheduler interprets memory values with the same binary units so a value means the same thing on both schedulers. Invalid `--memory` and `--memory-swap` values are now rejected by `resource:limit` and `resource:reserve`. Closes #9153.
This commit is contained in:
Jose Diaz-Gonzalez committed 2026-10-21 14:07:00 -04:00
1 parent fa0f031560
commit c8e28c2445
13 files changed
+382 -30

No files matched your search

+1 -1
View File
@@ -1293,7 +1293,7 @@ Two differences are worth noting:
The `k3s` scheduler supports a minimal list of resource _limits_ and _reservations_:
- cpu: is specified in number of CPUs a process can access.
- memory: should be specified with a suffix of `b` (bytes), `Ki` (kilobytes), `Mi` (megabytes), `Gi` (gigabytes). Default unit is `Mi` (megabytes).
- memory: should be specified with a suffix of `b` (bytes), `k` (kilobytes), `m` (megabytes), `g` (gigabytes), `t` (terabytes) or `p` (petabytes). Suffixes are case-insensitive and may be followed by `i` and/or `b`, such as `512Mi`, `1G`, or `1GiB`. Units are binary, so `1G` and `1Gi` are equivalent. Default unit is `Mi` (megabytes).
If unspecified for any task, the default reservation will be `.1` CPU and `128Mi` RAM, with no limit set for either CPU or RAM. This is to avoid issues with overscheduling pods on a cluster. To avoid issues, set more specific values for at least resource reservations. If unbounded utilization is desired, set CPU and Memory to `0m` and `0Mi`, respectively.