captainzmc commented on PR #3286: URL: https://github.com/apache/ozone/pull/3286#issuecomment-1092038544
> > also writing large keys is very unlikely to happen in practice. > > I don't agree with that at all. Large keys, small keys - they will all get written. I mean here is to write a big key when the quota threshold is near(a normal user can add a quota whenever a quota exception is raised). Unless, of course, the user intends to. > I agree there are problems with the current approach, especially needing to allocate in full block increments. However quota should stop someone from using more than their allocated space on the cluster. What if I start writing massive keys to effectively DDOS the cluster. I can just write and write beyond my quota TB's of data and never commit my keys. I can start many clients to do this in parallel and fill the entire cluster with garbage data. > In some respects, a failed to write key is using space and it could be argued it should count against the quota. It could be argued either way. However the ability for a user to write huge amounts of data beyond their quota, without the system blocking them in some way is bad I think. Agree, it's true that the new implementation does not prevent users with access from writing large keys over and over again. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
