sqitch-rebase - Revert and redeploy database changes
sqitch [options] rebase [<database>] sqitch [options] rebase [<database>] --onto-change <change> sqitch [options] rebase [<database>] --onto-change <change> --upto-change <change> sqitch [options] rebase [<database>] <change> sqitch [options] rebase [<database>] <change> --upto-change <change> sqitch [options] rebase [<database>] <change> <change>
Revert and redeploy changes to the database. It's effectively a shortcut for running sqitch revert
and sqitch deploy
in succession.
More specifically, starting from the current deployment state, changes will be reverted in reverse the order of application. All changes will be reverted unless a change is specified, either via --onto
or with no option flag, in which case changes will be reverted back to that change. If nothing needs to be reverted, a message will be emitted explaining why and nothing will be reverted.
Once the revert finishes, changes will be deployed starting from the deployed state through the rest of the deployment plan. They will run to the latest change in the plan, unless a change is specified, either via --upto
or with no option flag, in which case changes will be deployed up-to and including that change.
If the database has not been deployed to, or its state already matches the specified change, no reverts will be run. And if, at that point, the database is up-to-date, no deploys will be run.
The <database>
parameter specifies the database to which to connect, and may also be specified as the --target
option. It can be target name, a URI, an engine name, or plan file path.
-t
--target
The target database to which to connect. This option can be either a URI or the name of a target in the configuration.
--onto-change
--onto
Specify the reversion change. Defaults to reverting all changes. See sqitchchanges for the various ways in which changes can be specified.
--upto-change
--upto
Specify the deployment change. Defaults to the last point in the plan. See sqitchchanges for the various ways in which changes can be specified.
--mode
Specify the reversion mode to use in case of deploy failure. Possible values are:
all
In the event of failure, revert all deployed changes, back to --onto-change
. This is the default.
tag
In the event of failure, revert all deployed changes to the last successfully-applied tag. If no tags were applied, all changes will be reverted to --onto-change
.
change
In the event of failure, no changes will be reverted. This is on the assumption that a change is atomic, and thus may may be deployed again.
--verify
Verify each change by running its verify script, if there is one, immediate after deploying it. If a verify test fails, the deploy will be considered to have failed and the appropriate reversion will be carried out, depending on the value of --mode
.
--no-verify
Don't verify each change. This is the default.
-s
--set
Set a variable name and value for use by the database engine client, if it supports variables. The format must be name=value
, e.g., --set defuser='Homer Simpson'
. Overrides any values loaded from the deploy.variables
or revert.variables
configurations.
-d
--set-deploy
Set a variable name and value for use by the database engine client when deploying, if it supports variables. The format must be name=value
, e.g., --set defuser='Homer Simpson'
. Overrides any values from --set
or loaded from the deploy.variables
configuration.
-r
--set-revert
Sets a variable name to be used by the database engine client during when reverting, if it supports variables. The format must be name=value
, e.g., --set defuser='Homer Simpson'
. Overrides any values from --set
or loaded from the deploy.variables
and revert.variables
configurations.
--log-only
Log the changes as if they were deployed and reverted, but without actually running the deploy and revert scripts.
-y
Disable the prompt that normally asks whether or not to execute the revert.
[deploy.variables]
[revert.variables]
A section defining database client variables. The deploy.variables
configuration is read from the deploy
command configuration, on the assumption that the values will generally be the same. If they're not, use revert.variables
to override deploy.variables
.
These variables are Useful if your database engine supports variables in scripts, such as PostgreSQL's psql
variables, Vertica's vsql
variables, MySQL's user variables, and SQL*Plus's DEFINE
variables.
rebase.verify
deploy.verify
Boolean indicating whether or not to verify each change after deploying it.
rebase.mode
deploy.mode
Deploy mode. The supported values are the same as for the --mode
option.
[rebase.no_prompt]
[revert.no_prompt]
A boolean value indicating whether or not to disable the prompt before executing the revert. The rebase.no_prompt
variable takes precedence over revert.no_prompt
, and both may of course be overridden by -y
.
[rebase.prompt_accept]
[revert.prompt_accept]
A boolean value indicating whether default reply to the prompt before executing the revert should be "yes" or "no". The rebase.prompt_accept
variable takes precedence over revert.prompt_accept
, and both default to true, meaning to accept the revert by default.
Part of the sqitch suite.