# Updating the currently running system. # See https://github.com/rear/rear/issues/841 # Without a RECOVERY_UPDATE_URL there is nothing to do: test "$RECOVERY_UPDATE_URL" || return 0 # With a RECOVERY_UPDATE_URL ensure 'curl' is actually there # because that 'curl' was added to the default PROGS array # (see https://github.com/rear/rear/issues/1156) # is not sufficient to ensure 'curl' is actually there. # This test is run in particular during "rear mkbackup/mkrescue" # so that it errors out if 'curl' is required but not there: has_binary curl || Error "RECOVERY_UPDATE_URL requires that 'curl' is installed" # Currently updating ReaR is only supported during "rear recover" # because "rear recover" is run first and only once in the recovery system # (perhaps followed by several subsequent or simultaneous "rear restoreonly"). # Currently it is not tested what mess might happen for other workflows # that run many times in one same system like "rear mkbackup" in the normal system. # Furthermore it seems to make not much sense to update ReaR in the normal system # via this special built-in ReaR functionality because in the normal system # one can manually update ReaR as anything else. test "$WORKFLOW" != "recover" && return # The actual work: # Tell the user that the recovery system will be updated: LogPrint "Updating recovery system with the content from '$RECOVERY_UPDATE_URL':" # Download the tar.gz that contains the update files. local update_archive_filename="recovery-update.tar.gz" # "curl -f" does not fail reliable (there are occasions where non-successful response codes will slip through) # therefore the HTTP response code is written to stdout ('-w') so that it can be explicitly tested afterwards # ('-s' does not show progress meter or error messages but 'S' it makes it show an error message if it fails) # "curl --verbose" messages go to stderr so that they go to the "rear recover" log file: local http_response_code=$( curl $verbose -f -s -S -w "%{http_code}" -o /$update_archive_filename $RECOVERY_UPDATE_URL ) # Only HTTP response code 200 "OK" is what we want (cf. https://en.wikipedia.org/wiki/List_of_HTTP_status_codes): test "200" = "$http_response_code" || Error "curl '$RECOVERY_UPDATE_URL' failed with HTTP response code '$http_response_code'." # Install the downloaded tar.gz at the root directory '/' of the recovery system- # "tar --verbose" messages go to stdout so that they appear on the terminal where "rear recover" was started: pushd / tar $verbose -xf $update_archive_filename || Error "Updating recovery system via 'tar -xf /$update_archive_filename' failed." popd # Tell the user that recovery system update is done: LogPrint "Updated recovery system."