The lastUpdated time returned by healthz call should be the latest
lastUpdated time among the proxiers. Prior to this commit, if proxy
is unhealthy, the returned lastUpdated time was lastUpdated time
of the unhealthy proxier.
Signed-off-by: Daman Arora <aroradaman@gmail.com>
We want to be sure that the maximum number of pods per claim are actually
scheduled concurrently. Previously the test just made sure that they ran
eventually.
Running 256 pods only works on more than 2 nodes, so network-attached resources
have to be used. This is what the increased limit is meant for anyway. Because
of the tightened validation of node selectors in 1.32, the E2E test has to
use MatchExpressions because they allow listing node names.
The original limit of 32 seemed sufficient for a single GPU on a node. But for
shared non-local resources it is too low. For example, a ResourceClaim might be
used to allocate an interconnect channel that connects all pods of a workload
running on several different nodes, in which case the number of pods can be
considerably larger.
256 is high enough for currently planned systems. If we need something even
higher in the future, an alternative approach might be needed to avoid
scalability problems.
Normally, increasing such a limit would have to be done incrementally over two
releases. In this case we decided on
Slack (https://kubernetes.slack.com/archives/CJUQN3E4T/p1734593174791519) to
make an exception and apply this change to current master for 1.33 and backport
it to the next 1.32.x patch release for production usage.
This breaks downgrades to a 1.32 release without this change if there are
ResourceClaims with a number of consumers > 32 in ReservedFor. In practice,
this breakage is very unlikely because there are no workloads yet which need so
many consumers and such downgrades to a previous patch release are also
unlikely. Downgrades to 1.31 already weren't supported when using DRA v1beta1.
KubeProxy operates with a single health server and two proxies,
one for each IP family. The use of the term 'proxier' in the
types and functions within pkg/proxy/healthcheck can be
misleading, as it may suggest the existence of two health
servers, one for each IP family.
Signed-off-by: Daman Arora <aroradaman@gmail.com>
By passing the context, if available, down into the actual wait loop it becomes
possible to use contextual logging when handling a crash. That then logs more
information about which component encountered the problem (WithName e.g. in
kube-controller-manager) and what it was doing at the time (WithValues e.g. in
kube-scheduler).
