moxi tool needs fix (was: [2.0.2] memcachetest+moxi/mcsoda: downstream timeout)

know there are already a few bugs with the problem(CBSE-475/498...), but will create a new one, as no separate ticket for 2.0.2 and now there are automation tests to reproduce it

http://qa.hq.northscale.net/view/2.0.1/job/ubuntu-64-2.0-connections-tests
./testrunner -i /tmp/connections-tests.ini get-logs=True,wait_timeout=180,get-cbcollect-info=True -t connectionstests.ConnectionTests.multiple_connections_using_memcachetest

steps in test with 1 node cluster:
1. run moxi on node
nohup /opt/couchbase/bin/moxi -u root -Z usr=Administrator,pwd=password,port_listen=51500,concurrency=1024,wait_queue_timeout=200,connect_timeout=400,connect_max_errors=3,connect_retry_interval=30000,auth_timeout=100,downstream_conn_max=16,downstream_timeout=5000,cycle=200,default_bucket_name=default http://10.5.2.13:8091/pools/default/bucketsStreaming/default -d
2. run mcsoda through moxi
nohup python lib/perf_engines/mcsoda.py 10.5.2.13:51500 vbuckets=1024 doc-gen=0 doc-cache=0 ratio-creates=0.9 ratio-sets=0.9 min-value-size=1024 max-items=1000000 ratio-deletes=0.1 exit-after-creates=0 threads=4 prefix=default &
3.run memcachetest on the node
/tmp/memcachetest/memcachetest -h localhost:11211 -i 100000 -W 320 -t 320 -c 0 -M 2

result:

2013-04-12 08:33:37 | INFO | MainProcess | MainThread | [remote_util.log_command_output] ASCII set failure: SERVER_ERROR proxy downstream timeout
2013-04-12 08:33:37 | INFO | MainProcess | MainThread | [remote_util.log_command_output] Failed to set [29735]: during populate.

the same behaviour is for the test with 3 node. Tests can be expanded through parameters + can also be added additional scenarios as mentioned in http://hub.internal.couchbase.com/confluence/display/QA/QE+Test+Enhancements+for+2.0.2