rgw: Conditional PUT on ETag

[xylv]

Fixes: #8562

Signed-off-by: Ray Lv raylv@yahoo-inc.com

[xylv]

@yehudasa Thanks. That's great to see it. I'm going to squash a complete clean commit in this PR for you to review it.

[liewegas]

@xylv is this ready to review or is another commit coming?

[xylv]

@liewegas It's ready to review.

[liewegas]

I think @yehudasa needs to review. What would make this so much better would be a patch for s3-tests.git that tests the feature ... :)

[xylv]

@liewegas I agree that it's better to have the functional test cases. I'm looking into the s3-tests. The patch will be in a separated PR.

[loic-bot]

FAIL: make check on 5177821 output is http://paste.ubuntu.com/9450801/

Sent from GH.

[ghost]

@xylv are you able to run make check successfully on your local machine ? If so you can ignore loic-bot output and I'll investigate why it fails when you succeed. It is an experimental bot ;-)

[xylv]

@dachary I managed to run the make check on my local environment except for some cases on xfs. But it is as expected because I'm on a ext4 rather xfs filesystem. It's wired to see that *** DEVELOPER MODE: setting PATH, PYTHONPATH and LD_LIBRARY_PATH *** in the above log. I didn't see such messages in my local run. Just a guess, should that be related to the Python nose package is missing?

[ghost]

Just a guess, should that be related to the Python nose package is missing?

Python nose is installed on the machine running loic-bot make check. Could you add to http://paste.ubuntu.com/ the output of make check you have for comparison ?

[yehudasa]

@xylv @dachary could be that it's failing due to being based off an old branch, I don't see in the logs anything related to rgw

[ghost]

@yehudasa it could be the cause. @xylv do you mind rebasing on master to see if that makes a difference ?

[xylv]

@yehudasa @dachary I have rebased on master and run the make check again. Please find out the log at http://paste.ubuntu.com/9487775/ To test if rebasing works, also sent out the rebased PR: #3160