Decemeber: 17th

2024

• 02:26AM

  Kilobase

  During the migration process, we made a small mistake with the longhorn and now have to redeploy the database. I am going to take that time to see if there are any updates that we need to do as well.

• 03:20AM

  PvC

  We got a couple persistent volume claims that are stuck in a weird loop, where we are not too sure if they can be found or not. To help debug this porblem, we will be running these two commands:

  kubectl describe pvc pvc-1696b3cc-da05-4753-872c-f88f13e20d0a
  kubectl get pv | grep pvc-1696b3cc-da05-4753-872c-f88f13e20d0a
  

  Then it returns this:

  Error from server (NotFound): persistentvolumeclaims "pvc-1696b3cc-da05-4753-872c-f88f13e20d0a" not found
  8Gi        RWO            Delete           Terminating   armada/redis-data-redis-master-0                 longhorn       <unset>                          86d
  

  To grab the location for it, we went ahead and did this:

  kubectl edit pv pvc-1696b3cc-da05-4753-872c-f88f13e20d0a
  

  This will allow us to make a couple quick edits to the finalizers that hold that pvc. In this first case we want to just remove the redis data block, so we will strip out the finalizers holding it. Then do this:

  kubectl delete pv pvc-1696b3cc-da05-4753-872c-f88f13e20d0a --grace-period=0 --force
  

  We need to locate them inside of our longhorn.

  orphan-ee68325bd461018c0d1103776ad999b60263d56395c9f9364bc5bd7d2b08a844
  orphan-7bc0707b775aae99dc1927e76ac1677839a2ef7eacbb6509014aafe5cc81c77e
  

  The next option might be just best to just delete the deployment and start again?

  Okay we know the redis sits in the armada-naval, so we will go ahead and rebuild it from there. The first step will be the name, which is going to be armada-naval and it will deploy to the armada namespace. The repository url will be https://github.com/KBVE/kbve.git and the branch will be dev. Afterwards, we want to enable the self healing and set the paths to /migrations/kube/charts/armada. Hmm, there is another option to keep resources but we can let longhorn handle that for us. While that gets pushed up and prepares to build itself out, we can shift over to the AWS and prepare for the Kilobase launch.

• 03:49AM

  Kilobase

  Time to bring the kilobase back around but with more updates and minor fixes as well. We need to pull the s3 backup and do the cluster auto-migration / repair but I did want to deploy a new instance of kilobase, hmm. I will hold off on that rust adaption for now, too many problems can be an issue for us. Same as before but we will swap out the /migrations/kube/charts/kilobase path.

• 05:03PM

  Realtime

  Now I want to focus on getting the realtime supabase to work, hopefully without having too many errors in this process.

• 11:24PM

  Redis

  After getting the long horn and size situation resolved, we will be moving forward with the redis deployment. The goal of the redis deployment is to just get a better understanding of how it will work in our eco-system. Let me push up the phase zero of the deployment and see where it will go from there. Under Armada, I am going to have to add another deployment change.