Change the run.sh script to build the e2e executable and run the tests
using the executable.

This has several advantages:
- Document how the build an executable
- Test the built executable for every test run
- When running the executable the default -test.timeout is disable, so
  we don't need to specify it. This makes running the tests directly
  easier.

Signed-off-by: Nir Soffer <[email protected]>

To make it easier to inspect the log after failures. This is a quick
hack until we create the log properly (RamenDR#1722).

Signed-off-by: Nir Soffer <[email protected]>

Previously we uploaded only the gathered environment, but we really want
all data from the test when we try to debug a failure. Add the
e2e/ramen-e2e.log to the archive, and rename the archive to reflect the
new content.

Example content of the archive:

    % tree e2e.12932857944-1 -L3
    e2e.12932857944-1
    ├── e2e
    │   └── ramen-e2e.log
    └── test
        └── gather.rdr
            ├── dr1
            ├── dr2
            ├── gather.log
            └── hub

Signed-off-by: Nir Soffer <[email protected]>

Mainly for testing the previous changes, but this may be useful to keep
so we can debug "successful" build, or compare successful build with
failed build.

Gathering data is quick (~10 seconds), and the compressed tarball is
relatively small (7 MiB) and is automatically deleted after 15 days, so
always collecting data does not cost a lot.

Signed-off-by: Nir Soffer <[email protected]>