Our Gitlab v 14ee is hosted in GCP via helm, and we need to move the gitaly
volume to be hosted on a faster ssd gcp disk. I am following this guide
however I do not know which pods, statefulsets, or services depend on Gitaly being up, that can be scaled down or stopped to perform this disk data transfer.
When using Helm/Kubernetes, how does one effectively put the application into maintenance mode to do tasks like this?
Should we use kubectl
commands to scaled some services down? If so which ones?
The overall process to do this would be :
- set volumeReclaimPolicy to be
Retain
[done] - stop services ???
- take snapshot of gcp disk
- use snapshot to create new faster gcp disk
- update helm and pvc per Managing Persistent Volumes | GitLab
any help appreciate it ty!
% k get pods
NAME READY STATUS RESTARTS AGE
gitlab-runner-7947c948db-mx2q9 1/1 Running 0 21d
gl-gitaly-0 1/1 Running 0 21d
gl-gitlab-exporter-59d79cd9f7-9f5vh 1/1 Running 0 21d
gl-gitlab-pages-7f5588446d-k99q4 1/1 Running 0 21d
gl-gitlab-shell-c859f8cf6-8rs64 1/1 Running 0 21d
gl-gitlab-shell-c859f8cf6-hl6rk 1/1 Running 0 21d
gl-issuer-30-d4lz9 0/1 Completed 0 21d
gl-migrations-30-557nc 0/1 Completed 0 21d
gl-registry-595d74fc6d-7vk8q 1/1 Running 0 21d
gl-registry-595d74fc6d-qc8p2 1/1 Running 0 21d
gl-sidekiq-all-in-1-v2-8745df694-gngjd 1/1 Running 0 21d
gl-toolbox-7f5cf466fb-rt7xg 1/1 Running 0 21d
gl-webservice-default-767db659fb-ffhkg 2/2 Running 0 21d
gl-webservice-default-767db659fb-tnnw6 2/2 Running 0 21d