#1 Demo `webackup` fails on F28
Opened 3 years ago by netvor. Modified 3 years ago

Note: this was originally posted by @mbocek to github (previous forge where this project was hosted for a while):


jats-demo.noarch 0.0.0-1.fc28
shellfu-bash-xcase.noarch 0.0.1-1.fc28

$ jattool runtest /usr/share/jats/suite/demo/webackup
JAT__TEST_ID="jats://my.example.org/foodept/barpkg//webackup"
old_test_id="jats://redhat.com/oamg/preupgrade-assistant//preupg/result-files"
cannot reload session of different test

Workaround: $ rm -rf /var/tmp/jat/session/ to clean session from a previous failed test.


Thanks for the report!

This happens because of the way how JAT is designed to support interrupted test sessions (eg. reboot within test). The mechanism is kind of stupid: anytime when a session is being initiated but there is already an active one, it's automatically regarded as attempt to reload the old one.

At least it fails because test ids don't match. OTOH we have to expect tests to be able to crash, so I'll have to figure out how to handle this in more robust way.

Ideas:

  • just finish the old session, if found,
  • don't reload unless specified on CLI,
  • don't reload unless it's specified by the previous test (eg. via event in log or something),
  • OK, just re-think the whole reloading idea and add more proper support for it.

For now I'll try to at least provide some more useful message. Also a simple session management CLI that will allow user to investigate what's running and throw it away without having to know the /var/tmp path, etc.

Metadata Update from @netvor:
- Issue tagged with: changes--func, isa--bug

3 years ago

Login to comment on this ticket.

Metadata