Commit b65313a
Fixing Module CR deletion during node reboot
In case node it rebooted(for any reason) and the Module CR is being
deleted at the same period of time, the following sequence happens:
1) node becomes NotReady
2) module-nmc controller removes Spec from NMC
3) nmc controller does not schedule any unloader pod, since the node is
not ready
4) node becomes ready
5) nmc controller does not schedule unloader pod, since after node
reboot it ignores scheduling unloader pod, since the kernel module is
not loaded anyway
6) no unloader pod running, meaning Status in the NMC is not deleted,
meaning Module CR cannot be finalized, since Status in NMC is present
Solution:
When nmc controller processes Modules without Spec, and the node has
been rebooted, the nmc controller will delete the Status from NMC.1 parent efdbe15 commit b65313a
File tree
2 files changed
+12
-3
lines changed- internal/controllers
2 files changed
+12
-3
lines changed| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
408 | 408 | | |
409 | 409 | | |
410 | 410 | | |
411 | | - | |
412 | | - | |
| 411 | + | |
| 412 | + | |
| 413 | + | |
| 414 | + | |
413 | 415 | | |
414 | 416 | | |
415 | 417 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
793 | 793 | | |
794 | 794 | | |
795 | 795 | | |
| 796 | + | |
| 797 | + | |
796 | 798 | | |
797 | 799 | | |
798 | 800 | | |
| |||
801 | 803 | | |
802 | 804 | | |
803 | 805 | | |
| 806 | + | |
804 | 807 | | |
805 | 808 | | |
806 | 809 | | |
| |||
820 | 823 | | |
821 | 824 | | |
822 | 825 | | |
823 | | - | |
| 826 | + | |
| 827 | + | |
| 828 | + | |
| 829 | + | |
| 830 | + | |
824 | 831 | | |
825 | 832 | | |
826 | 833 | | |
| |||
0 commit comments