Nope, there are no timeouts, the queries are fast and 95% in cache,
this looks like a region server tried to read some memory buffer and
get 0 bytes in return.

On Mon, Nov 14, 2011 at 3:10 PM, Stack wrote:
On Mon, Nov 14, 2011 at 3:05 PM, Jack Levin wrote:
No custom code. and I did enable RPC logging to see what might be
wrong, but nothing is showing that would be considered an error.  Its
happening on two of our clusters that run different functions, one
uses thrift, and the other is REST.  From the looks of the stack trace
it seems like a low level java error.   One of our clusters runs scans
without filters, and the other is just PUT and GET, both get errors.
We can add some more debug code into the source and try if you can
suggest of a way to produce more debugging info.
What about other side of the connection? Is client going away on us?
Maybe these are slow queries and client has given up by the time the
server gets around to processing the request?


Search Discussions

Discussion Posts


Related Discussions

Discussion Navigation
viewthread | post
posts ‹ prev | 7 of 7 | next ›
Discussion Overview
groupuser @
categorieshbase, hadoop
postedNov 10, '11 at 3:24a
activeNov 15, '11 at 3:51a

3 users in discussion

Jack Levin: 4 posts Stack: 2 posts Ted Yu: 1 post



site design / logo © 2018 Grokbase