Migration tool documentation #1134
Conversation
Co-authored-by: Nikita Vaniasin <nikita.vanyasin@gmail.com>
…docs into DOC-189/migration-tool
| When starting the migration, specify the generated files on the command line: | ||
| | ||
| ```bash | ||
| --agent.cacert=tls.crt --agent.keyfile=tls.keyfile --agent.client-auth-cacert=client-auth-ca.crt --agent.client-auth-keyfile=client-auth.keyfile |
There was a problem hiding this comment.
what about reformat with \
how does this relates to the source.cacert and source.keyfile
as dummy I don't know what to do tbh
There was a problem hiding this comment.
source.* params are related to setting up connection to local cluster.
agent.* params are related to configuring local arangosync migration agent. These certificate options are not required.
| | ||
| An optional `--abort` option is supported. If specified, the `stop` command | ||
| will not check anymore if both deployments are in-sync and stops all | ||
| migration-related processes as soon as possible. |
There was a problem hiding this comment.
what happens with RO statuses in this case?
There was a problem hiding this comment.
@Robert-Stam
Source cluster will not be switched into read-only mode. The target will be switched back to default read-write
| --source.server-mode=readonly | ||
| ``` | ||
| The `--source.server-mode` option allows you to specify the desired server mode. | ||
| Allowed values are `readonly` or `default`. |
There was a problem hiding this comment.
What about RO/default mode for the Oasis deployment?
There was a problem hiding this comment.
@Robert-Stam for now it is not clear why user will want to switch the read only mode at target deployment manually. We can add such feature in next versions if needed.
| | ||
| ### Switching the local cluster to read-only mode | ||
| | ||
| The `arangosync-migration set-server-mode` command allows switching [read-only mode](../http/administration-and-monitoring.html#update-whether-or-not-a-server-is-in-read-only-mode) |
There was a problem hiding this comment.
This need to be done in rare situations only, right? what is the actual scenario we need to use this?
There was a problem hiding this comment.
@Robert-Stam
I see two scenarios where it can be useful (both described at the end of the doc):
- user can switch the source deployment into read-only mode before calling
stop- it will allow the synchronisation to catch up.stopdoes the same before actually stopping but has some timeouts (which may be exceeded if arangodb is experiencing a lot of writes for a long period). - In future, the
stopcommand will switch the source deployment into 'read-only' mode by default - it is currently not implemented. But user may want to switch it back to default read/write for his own reasons.
oasis/cloud-migration-tool.md Outdated
| You can use the `set-server-mode` | ||
| subcommand to switch it back to default, if needed. | ||
| In case something goes wrong during the migration, the `stop` command is not | ||
| switching the source data cluster into read-only mode. |
There was a problem hiding this comment.
How can I execute a 'dry-run', so a try to move everything into oasis without my source deployment does into RO mode. This can be good for an initial test phase. However I want to end-up with a non RO Oasis deployment as well, so I can execute test on that.
There was a problem hiding this comment.
@Robert-Stam - as already discussed before - 'dry-run' is not implemented for now, unfortunately. That is for next iterations. And target deployment of course will go into default read/write mode after stopping the sync.
5 participants
preview: https://deploy-preview-1134--zealous-morse-14392b.netlify.app/docs/3.10/oasis/cloud-migration-tool.html