Fix e2e scheduler_predicates.go to have 10 minute timeout instead of 2
minutes.
On an n1-standard-1 with only this test running, 2 minutes was enough
to create 189 pods out of the 300 desired. On slower VMs or with more
activity on the master (such as with parallel tests), 2 minutes is
nowhere near enough. This is a functional test, not a performance
test.
Showing
Please
register
or
sign in
to comment