-
Notifications
You must be signed in to change notification settings - Fork 195
Run collect script under ci #194
Run collect script under ci #194
Conversation
Run the data collection script to see if it works. Partially fixes kata-containers#193. Signed-off-by: James O. D. Hunt <[email protected]>
Add blank lines to make the teardown script clearer. Signed-off-by: James O. D. Hunt <[email protected]>
If the CI run fails, run the data collection script to help provide further clues on the failure. Fixes kata-containers#193. Signed-off-by: James O. D. Hunt <[email protected]>
e1bfb09
to
962e27f
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
General principle lgtm - one query...
@@ -71,3 +71,6 @@ sudo -E PATH="$PATH" bash -c "make test" | |||
|
|||
echo "INFO: Checking log files" | |||
check_log_files | |||
|
|||
echo "INFO: Running data collection script" | |||
sudo -E PATH="$PATH" chronic kata-collect-data.sh |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think you know what I'm going to say ... ;-)
Does this currently come out in the main CI log? Any chance you can redirect the output (at least stdout) to a known log file location - and then we can use that output in the teardown script, and save:
- injecting the rather large collect script output into the main console log
- having to run the script more than once?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Chronic discards all output unless the command fails so although there are 2 calls, this one won't bloat the console logs unless the script itself is broken (which we would want to know about).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
ah, indeed, I didn't spot the chronic
- mea culpa.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I know what you're saying, but the two calls are for different reasons - this one basically checks the script "works" (very simplistically of course), whereas the other call captures the script output if a PR fails (implicitly expecting the script to work).
That teardown script is a bit fragile imho as it has to cater for two separate scenarios so I'd be a little bit leery of adding more complexity to it by getting it to find the output of this script and potentially chop it into bits, etc.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
lgtm
Run the data collection script to see if it works.
If the CI run fails, run the data collection script to help provide further clues on the failure.
Fixes #193.
Signed-off-by: James O. D. Hunt [email protected]