Just some stuff that I'm too bored to write again and too lasy to develop into separate tool(code). DevOps Automation routines mostly.

Warning! Use it strictly on your own risk

• AWS
  • Update user password (cli is the last resort ex.)
  • Search for something in some trashcan S3 bucket. Second is iterate over buckets with threaded tool(s4cmd)
  • Find subnets that are short of IPs
  • Get all user keys. For blameless security investigation, you know
  • List EC2s + "plainification" with gron
• SSM tips
  • Get all instances with inventory without those in terminated state.
  • Connect to an instance by name
• K8S
  • See Events sorted by timestamp
  • Merge kubeconfigs
  • Add profile to kubeconfig, all of profiles all clusters
  • Saving private persistent volume
  • Label all nodes with their type and lifecycle, mark your cattle ;)
  • Debugging a distroless(ex.) pod
• MySQL
  • Kill somebody's sessions in RDS
  • Self-explanatory
  • Skip replication errors, read error-logs before skipping. It is important to understand what you are skipping.
• Other
  • Delete ALL topics in Kafka
  • Inframamp
  • KUBE-OPS-VIEW

AWS

Update user password (cli is the last resort ex.)

aws --profile PROFILE iam update-login-profile --user-name USER --password

Search for something in some trashcan S3 bucket. Second is iterate over buckets with threaded tool(s4cmd)

aws --profile PROFILE s3 ls s3://BUCKET/ --recursive | grep PATTERN
for b in $(s4cmd -r ls | grep ftp | awk '{print $4}') ; do s4cmd -r ls ${b}1465/ ; done

Find subnets that are short of IPs

aws --profile PROFILE ec2 describe-subnets --filters "Name=vpc-id,Values=VPC_ID" | jq '.Subnets[] | .SubnetId + "=" + "\(.AvailableIpAddressCount)"'

Get all user keys. For blameless security investigation, you know

for u in $(aws --profile PROFILE iam list-users  | jq ".Users[].UserName" --raw-output); do   aws --profile PROFILE iam list-access-keys --user $u | jq '.AccessKeyMetadata[] | .UserName + ":" + .AccessKeyId' ; done

List EC2s + "plainification" with gron

aws ec2 describe-instances | gron | grep 'MetadataOptions.HttpTokens = "optional"' 

SSM tips

Get all instances with inventory without those in terminated state.

Terminated instances sometimes are a problem. More about it.

aws ssm get-inventory --filters '[{"Key":"AWS:InstanceInformation.InstanceStatus","Values":["terminated"],"Type":"NotEqual"}]'

One more tip is to send huge jsons to gron. Grep Json.

Connect to an instance by name

aws ssm start-session --target INSTANCE_NAME

K8S

See Events sorted by timestamp

k get events --sort-by='.lastTimestamp'

Merge kubeconfigs

cp ~/.kube/config ~/.kube/config.bak && KUBECONFIG=~/.kube/config:./ok-cluster/ok-cluster-eks-a-cluster.kubeconfig kubectl config view --flatten > /tmp/config && mv /tmp/config ~/.kube/config

Add profile to kubeconfig, all of profiles all clusters

aws eks --profile ok-dev update-kubeconfig --name eks-terra --alias ok-eks
for p in $(aws configure list-profiles) ; do for c in $(aws eks --profile $p list-clusters | jq '.clusters[]' | tr -d '\"') ; do echo $p $c ; aws eks --profile $p update-kubeconfig --name $c --alias $p:$c ; done ; done

Saving private persistent volume

kubectl patch pv $PV_NAME_i -p \
  '{"spec":{"persistentVolumeReclaimPolicy":"Retain"}}'

Label all nodes with their type and lifecycle, mark your cattle ;)

for n in $(kubectl get nodes -o 'jsonpath={.items[*].metadata.name}') ; do lb="" ;  for a in $(kubectl label --list nodes $n | sort | grep -e NodeType -e lifecycle | cut -d= -f 2) ; do lb="${lb}$a" ; done ; kubectl label nodes $n node-role.kubernetes.io/$lb= ; done

Debugging a distroless(ex.) pod

kubectl debug -it POD --image=IMAGE_WITH_TOOLS --target=CONT --share-processes

MySQL

Kill somebody's sessions in RDS

SELECT CONCAT('CALL mysql.rds_kill(',id,');')
FROM information_schema.processlist
WHERE user='UGLY_BASTARD';

Self-explanatory

SHOW OPEN TABLES WHERE In_use > 0;
SHOW ENGINE INNODB STATUS;

Skip replication errors, read error-logs before skipping. It is important to understand what you are skipping.

CALL mysql.rds_skip_repl_error;

Other

Delete ALL topics in Kafka

export KFK=KFK_HOST
for t in $(./bin/kafka-topics.sh --bootstrap-server  $KFK:9092 --list) ; do ./bin/kafka-topics.sh --bootstrap-server $KFK:9092 --topic $t --delete ; done
#-||- --describe | grep 'ReplicationFactor:1' ; done

Inframamp

terraform state pull | inframap generate --connections=false | dot -Tpng > ~/Downloads/schema.png

KUBE-OPS-VIEW

(kubectl proxy --accept-hosts '.*' &) ; docker run -it -p 8080:8080 -e CLUSTERS=http://docker.for.mac.localhost:8001 hjacobs/kube-ops-view

Migrate Grafna DB from MySQL to PgSQL

podman run -it     --workdir $(pwd)/podman     --volume $(pwd)/podman:$(pwd)/podman:rw,Z     --rm ghcr.io/techouse/mysql-to-sqlite3:latest     --sqlite-file baz-prod.db     --mysql-user grafana     --mysql-password hehe     --sql-database prod_monitoring     --mysql-host MY_HOST --use-buffered-cursors

pg_dump --schema-only -h PG_HOST -U grafana -d grafana > schema.sql 
psql -h PG_HOST -U grafana -d grafana -f  schema.sql 

cat podman/f.load
load database
  from sqlite://baz-prod.db
  INTO postgresql://grafana:HEHE@PG_HOST/grafana?sslmode=prefer
  with data only, reset sequences, prefetch rows = 100
  set work_mem to '32MB', maintenance_work_mem to '256 MB';

podman run --rm -it  --workdir $(pwd)/podman  --volume $(pwd)/podman:$(pwd)/podman:rw,Z  dimitri/pgloader:latest   pgloader f.load