Migration

Migration can be done using OpenObserve binary as a CLI.

./openobserve -h
OpenObserve is an observability platform that allows you to capture, search, and analyze your logs, metrics, and traces.

Usage: openobserve [COMMAND]

Commands:
  reset               reset openobserve data
  import              import openobserve data
  export              export openobserve data
  view                view openobserve data
  init-dir            init openobserve data dir
  migrate-file-list   migrate file-list
  migrate-meta        migrate meta
  migrate-dashboards  migrate-dashboards
  delete-parquet      delete parquet files from s3 and file_list
  help                Print this message or the help of the given subcommand(s)

Options:
  -h, --help     Print help

You must use the migrate-meta command to migrate meta store.

./openobserve migrate-meta -h
migrate meta

Usage: openobserve migrate-meta --from <from> --to <to>

Options:
  -f, --from <from>  migrate from: sled, sqlite, etcd, mysql, postgresql
  -t, --to <to>      migrate to: sqlite, etcd, mysql, postgresql
  -h, --help         Print help

You must use migrate-file-list command to migrate the parquet file list

./openobserve migrate-file-list -h                         
migrate file-list

Usage: openobserve migrate-file-list [OPTIONS] --from <from> --to <to>

Options:
  -p, --prefix <prefix>  only migrate specified prefix, default is all
  -f, --from <from>      migrate from: sled, sqlite, etcd, mysql, postgresql
  -t, --to <to>          migrate to: sqlite, mysql, postgresql
  -h, --help             Print help

There are some environments for metadata store:

ZO_META_STORE: "postgres"
ZO_META_MYSQL_DSN: "mysql://user:password@server-address:3306/app"
ZO_META_POSTGRES_DSN: "postgres://user:password@server-address:5432/app"

Cluster mode migration

All the commands must be run in the pod compactor, If you can't login into the pod, please change your image tag to debug version, for example:

You are using v0.9.1 then you can set the image to debug version public.ecr.aws/zinclabs/openobserve:v0.9.1-debug

Please choose the version you are using and add suffix -debug to the tag, If you don't know which version you are using, you can login OpenObserve UI and click About menu to get it.

Change the image tag to debug version if you are running in kubernetes.
Deploy MySQL / PostgreSQL first, You also can create a RDS service in the cloud, or enable postgres using our helm chart:
```
postgres:
  enabled: true
```
You can copy the values from: https://github.com/openobserve/openobserve-helm-chart/blob/main/charts/openobserve/values.yaml#L521
Stop all the pods except compactor, you can scale all the other services to 0 and compactor to 1, Just keep only 1 compactor.
Login into compactor pod.

Set following environment variables depends on you want to use MySQL or PostgreSQL

$ export ZO_META_MYSQL_DSN="mysql://user:password@server-address:3306/app"
$ export ZO_META_POSTGRES_DSN="postgres://user:password@server-address:5432/app"

Run the command depends on you want to use MySQL or PostgreSQL

If there is no wal directory please create it.

mkdir -p data/openobserve/wal/

Remember change the data/openobserve/ to your real data directory.

./openobserve migrate-meta --from etcd --to mysql
./openobserve migrate-meta --from etcd --to postgres

This will migrate metadata.

./openobserve migrate-file-list --from etcd --to mysql
./openobserve migrate-file-list --from etcd --to postgres

This will migrate the file list.

If everything is fine then add these new environments to your deploy:

MySQL:

ZO_META_STORE: "mysql"
ZO_META_MYSQL_DSN: "mysql://user:password@server-address:3306/app"

PostgreSQL:

ZO_META_STORE: "postgres"
ZO_META_POSTGRES_DSN: "postgres://user:password@server-address:5432/app"

Restart OpenObserve, it should working with MySQL / PostgreSQL now.

You can connect to the database and you will see these tables:

+-----------------------+
| Tables_in_openobserve |
+-----------------------+
| file_list             |
| file_list_deleted     |
| meta                  |
| scheduled_jobs        | (>= v0.10.0)
| stream_stats          |
+-----------------------+
5 rows in set (0.00 sec)

Local mode migration

If you can't login into the pod, please change your image tag to debug version, for example: