|
|
+
keith@... 2012-11-07, 20:33
-
Re: Review Request: Patch submitted by Chris McCubbin to add thrift proxy to Accumulokeith@... 2012-11-07, 21:14
----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/7936/#review13220 ----------------------------------------------------------- /trunk/proxy/src/main/java/org/apache/accumulo/proxy/ProxyServer.java <https://reviews.apache.org/r/7936/#comment28444> An explicit NoSuchScanner exception may be useful. /trunk/proxy/src/main/java/org/apache/accumulo/proxy/ProxyServer.java <https://reviews.apache.org/r/7936/#comment28438> sync on batchScanner before checking if its null /trunk/proxy/src/main/java/org/apache/accumulo/proxy/ProxyServer.java <https://reviews.apache.org/r/7936/#comment28443> In past experience with Thrift throwing a TException in server side code woudl result in client seeing nothing or seeing something like an IOException instead of a TApplicationException. But this was a while ago, not sure what current behavior of thrift is. /trunk/proxy/src/main/java/org/apache/accumulo/proxy/ProxyServer.java <https://reviews.apache.org/r/7936/#comment28442> If user does not scan until end, then will scanner always be present in scannerMap and iteratorMap? If so this could lead to memory leaks. I think its a commom use case that users do not read all data from a scanner. /trunk/proxy/src/main/java/org/apache/accumulo/proxy/ProxyServer.java <https://reviews.apache.org/r/7936/#comment28445> would probably be more efficient to get the scanner once from the iteratorMap, rather than doing a lookup in the map for each loop iteration /trunk/proxy/src/main/thrift/proxy.thrift <https://reviews.apache.org/r/7936/#comment28441> can anything be done to make methods shorter? Is it possible to get rid of all of the tableOperation_ and instanceOperation_ prefixes? Does having different services (ie. AccumuloInstanceProxy, AccumloTableOpProxy) force us to run those services on separate ports? /trunk/proxy/src/main/thrift/proxy.thrift <https://reviews.apache.org/r/7936/#comment28440> the batch scanner will need to be closed in some way. I did not see this being done on the server side. The thrift API could provide a close method. Or the server side could time out idle batch scanner and close them. /trunk/proxy/src/main/thrift/proxy.thrift <https://reviews.apache.org/r/7936/#comment28439> Why not a list of Mutations? - kturner On Nov. 7, 2012, 8:33 p.m., kturner wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/7936/ > ----------------------------------------------------------- > > (Updated Nov. 7, 2012, 8:33 p.m.) > > > Review request for accumulo. > > > Description > ------- > > Patch submitted by Chris McCubbin to add thrift proxy to Accumulo > > > This addresses bug ACCUMULO-482. > https://issues.apache.org/jira/browse/ACCUMULO-482 > > > Diffs > ----- > > /trunk/pom.xml 1406772 > /trunk/proxy/README PRE-CREATION > /trunk/proxy/examples/python/README PRE-CREATION > /trunk/proxy/examples/python/TestClient.py PRE-CREATION > /trunk/proxy/examples/python/data/__init__.py PRE-CREATION > /trunk/proxy/examples/python/data/constants.py PRE-CREATION > /trunk/proxy/examples/python/data/ttypes.py PRE-CREATION > /trunk/proxy/examples/python/proxy/AccumuloProxy-remote PRE-CREATION > /trunk/proxy/examples/python/proxy/AccumuloProxy.py PRE-CREATION > /trunk/proxy/examples/python/proxy/__init__.py PRE-CREATION > /trunk/proxy/examples/python/proxy/constants.py PRE-CREATION > /trunk/proxy/examples/python/proxy/ttypes.py PRE-CREATION > /trunk/proxy/examples/ruby/README PRE-CREATION > /trunk/proxy/examples/ruby/accumulo_proxy.rb PRE-CREATION > /trunk/proxy/examples/ruby/data_constants.rb PRE-CREATION > /trunk/proxy/examples/ruby/data_types.rb PRE-CREATION > /trunk/proxy/examples/ruby/proxy_constants.rb PRE-CREATION +
keith@... 2012-11-07, 21:55
+
Chris McCubbin 2012-11-07, 21:40
+
Chris McCubbin 2012-11-07, 21:44
+
keith@... 2012-11-07, 21:41
+
Chris McCubbin 2012-11-07, 21:55
+
keith@... 2012-11-07, 22:05
+
keith@... 2012-11-07, 21:44
|