Alexey Serbin has posted comments on this change.

Change subject: [c++client] added TestAutoFlushBackgroundDropSession

Patch Set 1:

File src/kudu/client/

PS1, Line 2431: its pending data should not be flushed.
> Are you sure these are the semantics we want? They make sense for MANUAL_FL
Great observation!

We don't have requirements for that, but after some consideration I think it's 
worth discussing this, indeed.

>From one hand, the behavior you mentioned would be more user-friendly and 
>elegant from the coding perspective: knowing the session is run in 
>AUTO_FLUSH_BACKGROUND mode it's not necessary to call Flush() in the very end 
>of the 'logical' session.

>From the other hand, changing the current behavior would  require providing a 
>means to report an error encountered during flushing the reminding operations, 
>right?  Otherwise, how could a user know that the operations of the abandoned 
>session were flushed successfully?

That's said, I would consider declaring this a usability bug  and addressing 
this issue in a separate changelist.

Would it be OK?

PS1, Line 2433: due
> Nit: due to

Line 2446:   SleepFor(MonoDelta::FromMilliseconds(3 * kFlushIntervalMs));
> warning: either cast from 'int' to 'int64_t' (aka 'long') is ineffective, o

To view, visit
To unsubscribe, visit

Gerrit-MessageType: comment
Gerrit-Change-Id: If136872a4200e18acbb587276fed28975afe3810
Gerrit-PatchSet: 1
Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-Owner: Alexey Serbin <>
Gerrit-Reviewer: Adar Dembo <>
Gerrit-Reviewer: Alexey Serbin <>
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Tidy Bot
Gerrit-HasComments: Yes

Reply via email to