How big is too big for a DistributedCache file hadoop? -
are there guidelines whether distribute file using distributed cache or not ?
i have file of size 86746785 (i use hadoop dfs -dus
- don't know if in bytes
or what). idea distribue file ?
the viable answer "it depends".
what have consider using distributed cache file gets copied every node involved in task, takes bandwidth. also, if want file in distributed cache, you'll keep file in memory, you'd have take consideration.
as case -- yes, bytes. size 86 mb, fine distributed cache. within couple hundred mbs should still be.
Comments
Post a Comment