Closed ZihanJiang96 closed 1 year ago
the CI failed, but i don't have permission to rerun it
@ZihanJiang96 nice work!
Just two concerns:
Merging #351 (343f99d) into master (fc20675) will increase coverage by
5.95%
. The diff coverage is54.05%
.
@@ Coverage Diff @@
## master #351 +/- ##
==========================================
+ Coverage 39.22% 45.17% +5.95%
==========================================
Files 7 7
Lines 724 757 +33
==========================================
+ Hits 284 342 +58
+ Misses 415 376 -39
- Partials 25 39 +14
Flag | Coverage Δ | |
---|---|---|
unittests | 45.17% <54.05%> (+5.95%) |
:arrow_up: |
Flags with carried forward coverage won't be shown. Click here to find out more.
Impacted Files | Coverage Δ | |
---|---|---|
controllers/rollingupgrade_controller.go | 35.71% <42.85%> (+35.71%) |
:arrow_up: |
controllers/upgrade.go | 46.07% <66.66%> (+0.16%) |
:arrow_up: |
controllers/common/utils.go | 31.57% <100.00%> (+31.57%) |
:arrow_up: |
:mega: We’re building smart automated test selection to slash your CI/CD build times. Learn more
CI green, let me add some unit test to improve test coverage
problem currently
max-parallel
only controls the number of concurrent reconcile() func, in a short period of time, all the IGs will enter the reconcile() at least once, and start to spin up new nodes / drain the old nodes. This will cause a memory spike in upgrade-manager.proposal
max-parallel
should limit the max number of RollingUpgrade CRs inrunning
state.changes introduced by this PR
max-parallel
number of IGs in theAdmissionMap
.AdmissionMap
is zero, then possibly the pod restarted, fetch all CRs in running state by label, add to admission map, continue the reconcilation against these running IGs.tests
max-parallel
to 10, the memory consumption is 310MBmax-parallel
to 10, the memory consumption is 700MB, the upgrade time is 85min