tgravescs commented on issue #25299: [SPARK-27651][Core] Avoid the network when
shuffle blocks are fetched from the same host
URL: https://github.com/apache/spark/pull/25299#issuecomment-558642474
test this please
This is an
tgravescs commented on issue #25299: [SPARK-27651][Core] Avoid the network when
shuffle blocks are fetched from the same host
URL: https://github.com/apache/spark/pull/25299#issuecomment-558299984
Definitely agree, sounds like gain and been wanting this for a while so
thanks @attilapiros
tgravescs commented on issue #25299: [SPARK-27651][Core] Avoid the network when
shuffle blocks are fetched from the same host
URL: https://github.com/apache/spark/pull/25299#issuecomment-558228346
ok that makes sense. thanks. So it seems we fetch all the local blocks in
first and then
tgravescs commented on issue #25299: [SPARK-27651][Core] Avoid the network when
shuffle blocks are fetched from the same host
URL: https://github.com/apache/spark/pull/25299#issuecomment-558202319
I only skimmed this but overall looks good, like this approach.
When fetching the host
tgravescs commented on issue #25299: [SPARK-27651][Core] Avoid the network when
shuffle blocks are fetched from the same host
URL: https://github.com/apache/spark/pull/25299#issuecomment-527909451
So the downside to using a disk is that at least on yarn containers may not
see the same
tgravescs commented on issue #25299: [SPARK-27651][Core] Avoid the network when
shuffle blocks are fetched from the same host
URL: https://github.com/apache/spark/pull/25299#issuecomment-517397877
yes we had talked about this and I looked briefly, but like mentioned in the
description you