Tim Armstrong has posted comments on this change.

Change subject: IMPALA-5113: fix dirty unpinned invariant
......................................................................


Patch Set 3:

(5 comments)

http://gerrit.cloudera.org:8080/#/c/6469/3/be/src/runtime/bufferpool/buffer-pool.cc
File be/src/runtime/bufferpool/buffer-pool.cc:

PS3, Line 424: Released reservation and dirty unpinned should cancel out.
> I'm not sure what this comment is saying.
Removed.


PS3, Line 517: reservation
> why do we pass this rather than just get it from the client?
Historical reasons - it used to be part of the ClientHandle rather than the 
Client. I missed the potential cleanup. Fixed this and a couple of other places


PS3, Line 520: the invariant
> the eviction policy.
Done


Line 541:   // We cleared up enough space to bump 'buffers_in_use_bytes_'.
> this comment is kind of cryptic. what is it trying to tell me?
Removed the code and comment as part of the PageList refactor.


Line 542:   buffers_in_use_bytes_ += len;
> It's starting to become difficult to keep track of what these aggregators a
Went with a PageList wrapper. It reduces the amount of bookkeeping in complex 
parts of the code so seems like a win. I had to move the Page struct definition 
up to the top of the -internal.h file to get the inline functions to compile.

I also ended up changing the Remove() function to return a bool (since it 
simplified the wrapper). This led to a cascade of other changes where I could 
change the Contains()/Remove() pattern to a single Remove() call.


-- 
To view, visit http://gerrit.cloudera.org:8080/6469
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I07e08acb6cf6839bfccbd09258c093b1c8252b25
Gerrit-PatchSet: 3
Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-Owner: Tim Armstrong <[email protected]>
Gerrit-Reviewer: Dan Hecht <[email protected]>
Gerrit-Reviewer: Tim Armstrong <[email protected]>
Gerrit-HasComments: Yes

Reply via email to