Thanks For such an informative reply. I have already implemented s3a:// and yes only is the solution.  The other one,i.e. changing to /tmp dir is an intelligent workaround.     ... View more

> So, to generalize, the mechanism level subcodes can always be taken as some failure in communicating with KDC, right? Yes, it can be always taken as something wrong in the Kerberos layer (not necessarily only KDC, could also be things such as bad enctypes in keytab, etc., but always Kerberos mechanism related) > I also see that despite this error, ZK does continue to function ... so is this error to be really treated seriously? Did a retry of the auth perhaps succeed? Its not normal for it to repeat the errors. ... View more

I'm running into something similar.  I'm on 5.4.2 building tables with have then analyzing with Impala and I get the same warnings, although the queries execute ok. Can you please share with me what you scripted to make "when one partition is always less than 800MB I set the block size for this table to 1GB" as you mention in your post? ... View more

Thanks Harsh, yes that is helpful. ... View more

Thanks for the input  I am suspecting that it may be the region settings I am using that prematurely splits regions contributing to the latency Originally I set the regions at default to 10 gb, shrank it to 4 gb did a compaction and flush, the first writes where trying to reorg the regions  which is stemming from my first problem  http://community.cloudera.com/t5/Storage-Random-Access-HDFS/hbase-closed-upon-write/m-p/34001#U34001 I then increased the regions to 4.9 gb seems to be fine now The reason I am staying under 5 gb, is so that I can transfer to S3 which has a file size limit   ... View more

The only way I can think of to do that is to spawn a whole new application, impersonating the user, from within your AM. That may work, although I've never tried it. ... View more