ivandika3 commented on PR #1362:
URL: https://github.com/apache/ratis/pull/1362#issuecomment-4044489677

   > Yes, the client has to keep track of its seenIndex. They are similar in 
the sense the server has to wait for a particular index before serving a read:
   
   I see this would be similar idea as the stateID idea in HDFS Observer read, 
but I hope that we can have a completely stateless solution so we don't need to 
make client keep track of any index since it's brittle in distributed 
environment. In the case of Ozone, if we setup multiple S3Gs (each with their 
own client) under a reverse proxy, each S3 request from the same "client" can 
be routed to one of the S3G, each with its own client. If one client has seen 
index x, but the other client has seen x - 10, the issue might still happen. 
Similar idea is for distributed jobs where the executor and workers have 
different clients. Please correct me if I'm misunderstood.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to