Previously, we'd check for the log device (passed with --log-to) only
after the user confirmed to proceed with the migration. And then, if the
device was not valid, takeover would keep running without storing the
stage 2 logs anywhere.
In other words, a mistake in the command-line arguments could cause
important logs to be lost.
With this commit, we check the the log device at a much earlier point
(before asking the user for confirmation to proceed), and if the device
provided is no good, we log an error message and exit. This allows the
user to fix the command-line arguments and run again.
Previously, we'd check for the log device (passed with
--log-to
) only after the user confirmed to proceed with the migration. And then, if the device was not valid, takeover would keep running without storing the stage 2 logs anywhere.In other words, a mistake in the command-line arguments could cause important logs to be lost.
With this commit, we check the the log device at a much earlier point (before asking the user for confirmation to proceed), and if the device provided is no good, we log an error message and exit. This allows the user to fix the command-line arguments and run again.