[
https://issues.apache.org/jira/browse/FLINK-3322?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15470419#comment-15470419
]
ramkrishna.s.vasudevan commented on FLINK-3322:
-----------------------------------------------
I am able to handle cases for the iterative jobs. But when we go to the
LargeRecordHandlers then it is much more trickier. Checking that part Will get
back on that.
Currently the design is that the MemoryAllocator will be passed on to the
Sorters and the memory allocator will have pre created memory segments. If the
memory allocator is created by Iterative tasks then we ensure that such
segments are not directly released to memory manager and retain them till the
iterative tasks receive termination signal.
In normal batch task cases - the memory allocators created are not to be kept
for further iterations and hence we close them out.
The sorters create read buffers, write buffers and large buffers. These are all
static based. But inside large record handler we have some dynamic way to
decide the number of records needed for keys and records.
Will get back on this. Any suggestions/feedbacks here?
> MemoryManager creates too much GC pressure with iterative jobs
> --------------------------------------------------------------
>
> Key: FLINK-3322
> URL: https://issues.apache.org/jira/browse/FLINK-3322
> Project: Flink
> Issue Type: Bug
> Components: Local Runtime
> Affects Versions: 1.0.0
> Reporter: Gabor Gevay
> Priority: Critical
> Fix For: 1.0.0
>
> Attachments: FLINK-3322.docx
>
>
> When taskmanager.memory.preallocate is false (the default), released memory
> segments are not added to a pool, but the GC is expected to take care of
> them. This puts too much pressure on the GC with iterative jobs, where the
> operators reallocate all memory at every superstep.
> See the following discussion on the mailing list:
> http://apache-flink-mailing-list-archive.1008284.n3.nabble.com/Memory-manager-behavior-in-iterative-jobs-tt10066.html
> Reproducing the issue:
> https://github.com/ggevay/flink/tree/MemoryManager-crazy-gc
> The class to start is malom.Solver. If you increase the memory given to the
> JVM from 1 to 50 GB, performance gradually degrades by more than 10 times.
> (It will generate some lookuptables to /tmp on first run for a few minutes.)
> (I think the slowdown might also depend somewhat on
> taskmanager.memory.fraction, because more unused non-managed memory results
> in rarer GCs.)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
