ODF documents, unless they are flat XML document (.fodt, .fods, ...) are zip containers with several .xml streams (content.xml, styles.xml, images, ...). Any capable (un)zip implementation should be able to detect the CRC errors I mentioned there. To inspect streams just unzip the document that already gives the errors I mentioned, zip integrity can also be checked by the InfoZip implementation's unzip -t filename.ods (for example) command.