This message was deleted.
# k3sa
adamant-kite-43734
02/13/2024, 4:53 PMThis message was deleted.
l
late-needle-80860
02/16/2024, 6:03 PMThe official etcd docs apply for k3s usage as well.
So read up up on the iops requirements of etcd in there. Use at minimum ssd type disks … preferably nvme.
Do not have regular workloads on the same nodes as etcd…consider splitting out etcd if you’re running embedded mode today. So that etcd is running on its own isolated setup.
a
abundant-lion-58619
02/16/2024, 6:09 PMHello, thanks for your attention!
Actually, I meet all that requirements (NVMe etc) except that the ping is higher and this is expected in my case. I've, of course, read the etcd docs and applied parameters for heartbeat timeout, therefore the system itself works stable, but as I understood, a 100ms timeout is hardcoded for warnings and therefore my syslog is growing very fast. Actually, I have to clean it daily and all other entries just dissappear between etcd warnings.
abundant-lion-58619
02/16/2024, 6:14 PMthere is some experimental feature for latest etcd to configure warning threshold, but I can't figure out how to apply it to k3s, because k3s has embedded etcd and I don't know how to access it. I even forgot how did I apply heartbeat timeout (which is 100% applied because I see this setting). When I try to apply this setting again on the node with
etcdabundant-lion-58619
02/16/2024, 6:22 PMI found that this issue is discussed on github but seems no solution till now 😞
abundant-lion-58619
02/16/2024, 6:25 PMI would change the logging level to ERROR or CRITICAL but don't know how to do that
abundant-lion-58619
02/16/2024, 6:29 PMI tried to change logging level for etcd according to official documentation, but it fails:
Copy code
$ etcd --log-level error
{"level":"fatal","ts":"2024-02-16T19:27:24.531+0100","caller":"etcdmain/etcd.go:204","msg":"discovery failed","error":"listen tcp 127.0.0.1:2380: bind: address already in use","stacktrace":"<http://go.etcd.io/etcd/server/v3/etcdmain.startEtcdOrProxyV2|go.etcd.io/etcd/server/v3/etcdmain.startEtcdOrProxyV2>\n\t/tmp/etcd-release-3.5.5/etcd/release/etcd/server/etcdmain/etcd.go:204\ngo.etcd.io/etcd/server/v3/etcdmain.Main\n\t/tmp/etcd-release-3.5.5/etcd/release/etcd/server/etcdmain/main.go:40\nmain.main\n\t/tmp/etcd-release-3.5.5/etcd/release/etcd/server/main.go:32\nruntime.main\n\t/usr/local/google/home/siarkowicz/.gvm/gos/go1.16.15/src/runtime/proc.go:225"}n
nutritious-tomato-14686
02/20/2024, 6:30 PMYou can't modify the embedded k3s etcd with the external etcd CLI. You need to pass these arguments directly when launching K3s using the
--etcd-argnutritious-tomato-14686
02/20/2024, 6:30 PM47 Views