etcd reading member-k0-key.pem every second (again) #19505
Labels
priority/awaiting-more-evidence
Lowest priority. Possibly useful, but not yet enough support to actually get it done.
stage/triaged
type/bug
Bug report criteria
What happened?
I have observed that etcd is reading the private key from disk continuously. The issue observed when an audit rule has been applied to catch the private keys being read.
The issue was discovered running a kubernetes cluster with three nodes deployed with kubespray in three virtual machines managed by vagrant.
What did you expect to happen?
etcd should read the key at start up and keep it in memory for further use.
My previous ticket was closed without any chance of justifying my point of view.
Here goes the replica:
Supporting dynamically loading the certificate does not imply on reading it from disk at every call.
The certificate can be held for a specific amount of time, and replaced when that time is expired. That could be even configurable.
Reading the file at every call can cause a bottleneck and harm the performance.
How can we reproduce it (as minimally and precisely as possible)?
On a k8s control plane node install auditd.
Create an audit rules for monitoring the key file
Inspect audit logs
Anything else we need to know?
No response
Etcd version (please run commands below)
Etcd configuration (command line flags or environment variables)
Etcd debug information (please run commands below, feel free to obfuscate the IP address or FQDN in the output)
Relevant log output
The text was updated successfully, but these errors were encountered: