unixfox / k8s Goto Github PK

encrypt nextcloud data

stash with restic through rclone and rclone with a lot of remotes

start active mysql backup servers of oracle managed DB.

help:

• https://lefred.be/content/master-slave-replication-with-mysql-8-0-in-2-mins/
• https://www.digitalocean.com/community/tutorials/how-to-set-up-master-slave-replication-in-mysql

add a slave on a live database:

• https://www.schalley.eu/2017/01/16/setting-up-mysql-replication-with-existing-data-master-slave/

https://coreos.com/flannel/docs/latest/kubernetes.html

kubectl annotate node name flannel.alpha.coreos.com/public-ip-overwrite=9.9.9.9

Move k3s datastorage for the API to a more reliable location.

switch to caddy but only for handling HTTPS and HTTP. still use nginx as a way to route http requests and handling rules
use on demand TLS as a way to avoid hardcoding any new domain in a config: https://caddyserver.com/docs/automatic-https#on-demand-tls

/24 is too small

cloudprovider: https://github.com/kubernetes/autoscaler/tree/master/cluster-autoscaler/cloudprovider/azure
helm chart: https://github.com/kubernetes/autoscaler/tree/master/charts/cluster-autoscaler

https://github.com/inlets/inlets-operator
https://github.com/inlets/inletsctl

be alerted if stash fails

rancher/system-upgrade-controller#19
rancher/system-upgrade-controller#23
rancher/system-upgrade-controller#22
https://rancher.com/docs/k3s/latest/en/upgrades/automated/

https://rancher.com/docs/k3s/latest/en/upgrades/automated/

make node.kubernetes.io/not-ready shorter

packages:

• wireguard-tools
• wireguard-dkms
• nfs-common

configure system:

• swap
• set external IPs for koli and node-external-ip

help for building that: https://raw.githubusercontent.com/squat/modulus/master/wireguard/daemonset.yaml

oracle: https://github.com/oracle/oci-cloud-controller-manager
alibaba: https://github.com/kubernetes/cloud-provider-alibaba-cloud
azure: https://github.com/kubernetes-sigs/cloud-provider-azure

solution for replicating mariadb using only local storage

taint the nodes from a config file instead of commands

keydb with s3 backend

discord support for botkube was added: kubeshop/botkube#268

https://github.com/roboll/helmfile#separating-helmfileyaml-into-multiple-independent-files

k3s-io/k3s#2400

https://github.com/kubernetes/autoscaler/tree/master/vertical-pod-autoscaler

• portwox
• longhorn
• storageos

autoscaler from/to zero node doesn't work with cluster-api, write a similar autoscaler for that using annotations.

Sometimes the connection doesn't work anymore between two wireguard connections

it is slow

one pod running with an ipv6 tunnel and synced with it.
this pod is attached to a created load balancer
define a low noschedule for a low downtime

oracle cloud, scaleway and aruba iso

• https://github.com/ctrox/csi-s3
• https://github.com/informatics-lab/s3-fuse-flex-volume
• https://github.com/zee-ahmed/kube-rclone

oracle cloud s3: https://blogs.oracle.com/linux/using-rclone-to-copy-data-in-and-out-of-oracle-cloud-object-storage

https://github.com/porter-dev/porter

https://github.com/projectatomic/docker/blob/master/docs/userguide/networking/default_network/ipv6.md

keel-hq/keel#449
https://github.com/micw/kube-autoupdate
https://github.com/arnehilmann/k8s-auto-updater

redeploy the cluster using external database for k3s?

have a sidecar in NGINX pod that inject a tunnel from a tunnelbroker
use a different tunnel name than he-ipv6!
schedule nginx only on the node where oci-ha-floating-ip is scheduled
switch nginx to deployment type
low node.kubernetes.io/not-ready

https://github.com/deislabs/osiris

could be useful for searx

right now if something fail it silently fail
netdata cloud + kubernetes integration ?
kubewatch alternative
zabbix?

https://github.com/kubermatic/machine-controller

• adapt for k3s
• change image for DO
• add compatibility with AWS access token
• change user data

for aws spot: use a termination handler so that the node gets evicted before it is terminated

annotations:
  rollme: {{ randAlphaNum 5 | quote }}

is not a good idea

https://gist.github.com/stewartpark/14a5690afbf717be04857d2cda1d8620

store files in a localpath? or using rclone mount from an oracle object storage?

stash is now installed but still considering duplicati

https://github.com/kubernetes/kubernetes/blob/master/cluster/addons/dns/nodelocaldns/nodelocaldns.yaml

helm charts:

• https://github.com/valerius257/k8s-nodelocaldns-helm
• https://github.com/lablabs/k8s-nodelocaldns-helm

related: squat/kilo#49

with helm charts museum in serverless?