Pig does not depends on MR historyserver directly, but MR framework does.
Pig needs to pull counter/diagnose information for each MR job it submits
from AM, but if AM dies, MR client will redirect the request to history
server. Note in Pig 0.13 and prior, there is a bug which keep Pig trying
histor
Has anyone experience this ? Thanks
On Fri, Jan 16, 2015 at 7:28 PM, canan chen wrote:
> I try to run pig on mapreduce, but would meet the following error if I
> didn't start mr history server.
> But I can run hive on mapreduce without MR historyserver, not sure why pig
> on mapreduce on MR hist
I try to run pig on mapreduce, but would meet the following error if I
didn't start mr history server.
But I can run hive on mapreduce without MR historyserver, not sure why pig
on mapreduce on MR historyserver. As my understanding historyserver is used
for recovery, so it should be OK without hist