lance

@xiaods as u can see the metrics , my disk is ssd and etcd metric "etcd_disk_wal_fsync" shows ok，and this problem only occurs when an old member(its termId is greater then...

i do not think so, as the etcd log shows ，when leader changes，the master just reply slower than the worker, and the leader change has already finished

> /lgtm I do not have the permission to change the jira status, could u please help to change the jiva to valid state

/cc @pacevedom

> Hey @lance5890, thanks for your contribution. > > Could you explain how increasing time between restarts will help you? What problem are you observing? > > /label ok-to-test We...

/retest-required

> I noticed you've taken steps to address the problem with StartLimitIntervalSec. However, I'm afraid that this change needlessly increases the timeout that some feature depend on and I'm hesitant...

found another container stuck in removing mount point ``` Apr 02 22:29:09 sphere75-node-3 crio[17278]: time="2025-04-02 22:29:09.059456865+08:00" level=info msg="Creating container: ccos-cluster-storage-operator/csi-snapshot-controller-5689576cd8-pmmkd/snapshot-controller" id=bd0c28fa-a8e9-46d4-8fa1-7500320aae24 name=/runtime.v1.RuntimeService/CreateContainer Apr 02 22:29:09 sphere75-node-3 crio[17278]: time="2025-04-02 22:29:09.215492912+08:00" level=error...

@odinuge could u please review this PR? /cc @odinuge

yes ,as memoryCapacity is calculated by the GetMachineMemoryCapacity in cadvisor , which get all the memory in /proc/meminfo, this contains the hugepage memory, I think it is resonable to caculate...