-
Notifications
You must be signed in to change notification settings - Fork 557
PYTHON-1232: protocol v5 out of beta #1099
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
@beobal so with those changes, things look good for C* <4. However, I'm having some failures with a secure 4.0 cluster. There is probably something missing in the driver. If I send bad creds, I can see that the auth failed response is not wrapped (checksumming). When sending good creds, it seems that the auth succeed response is wrapped, is that correct? |
@aboudreault that doesn't sound right, any response sent from server to client after the server has received the |
@aboudreault it's a driver issue, looks like I missed enabling checksumming for v5 in the authentication case. This patch fixes it:
|
That is indeed something I have already tried. I get a CrcMismatchException when enabling checksumming just before sending the auth message. I will retest everything with the CASSANDRA-14973's branch today. |
Interesting, I definitely see an error without that patch, but not with it. When you say "just before sending the auth message", do you mean the |
I just push your change in this branch. So with this driver branch and this C* branch: https://github.com/beobal/cassandra/commits/14973-trunk I create a cluster with ccm and this config:
I'm getting this error:
which is raised by the server:
Were you testing the same auth configuration? |
I wasn't setting the authorizer, but that shouldn't matter anyway as it's not used during authentication. Either way, I've followed your steps and created a single node ccm cluster from that branch, with those settings. Using this driver branch I see this:
The fact that it's your server logging the CRC error means the client isn't properly encoding the request. Am I doing something different here? |
Also, just to be sure, I verified that my client is actually connecting over v5:
|
looks like we are doing the same thing then. that must be something in my env... I will continue to debug to understand what's going on.. |
Ok. I found the issue in the driver. You can also reproduce it if you install the |
nice catch! confirmed I can repro with compression enabled. |
c81ca0d
to
4a00c22
Compare
Merging. Thanks! |
No description provided.