Restore Nimbus Back To Life With Better Docs && Backups

[mfekadu]

Objective

Use our current opportunity to improve documentation && reliability of Nimbus

Key Result

• Run through the MySQL setup documentation to deploy the MySQL server.
  • see if anyone has manual backups on their local machines
• Discuss && consider updating the MySQL setup docs
• Recreate the API VM instance on GCP
• Set up snapshots on GCP to promote #reliability

Details

• ??

[mfekadu]

Discuss && consider updating the MySQL setup docs

notes from MySQL setup docs in Wiki

1. MySQL is taught to Cal Poly Computer Science students in CSC 365
2. MySQL is reportedly used by large software companies, which suggests it is reliable, scalable, and maintainable
3. MySQL's relational model of storing data allows it to represent the relations between relevant objects. See the database schema
   (e.g., Dr. Khosmood's entry in the Professors table might have a relation to many entries in the Courses table and vice-versa. This is known as a many-to-many relation.)

[mfekadu]

Set up snapshots on GCP to promote #reliability

https://cloud.google.com/compute/docs/disks/create-snapshots

[snekiam]

Recap of what got us here:

Our instances on GCP had 'test' in their names, and we didn't setup snapshots. While cleaning up some instances in our GCloud account, our 'test' instances were removed, since there was confusion about which instances were actually being used. Since we had no snapshots, our database was gone.

We should setup our infrastructure so if an instance does get deleted, it does not cause data loss, and only causes us to need to recreate the instance and restore a backup.

[mfekadu]

Run through the MySQL setup documentation to deploy the MySQL server.

We are now considering CloudSQL instead of the Compute Engine way of managing MySQL as seen in the MySQL wiki doc

[snekiam]

Info on cloud sql: https://stackoverflow.com/a/34030484/13291759

[mfekadu]

good reasons for CloudSQL

• utf8mb4 supports emojis 😄 - @Jason-Ku

[snekiam]

How far did we get last week?

[snekiam]

The database is restored! We're working on geting the API back up now. GCP seems to be running backups now automatically.