plum-11-26-2012

plum-11-26-2012.txt

11 node cluster
running this workload now for the next 10 hours , doc size is 512 , i think this is going to push the cluster to dgm

{
    "name" : "plum",
    "desc" : "very simple key-value test",
    "loop" : true,
    "phases" : {
                "0" :
                {  
                    "name" : "simple_load",
                    "desc" :  "load items at 1k ops",
                    "workload" : "s:70,g:29,d:1,ccq:simplekeys,ops:8000",
                    "template" : "default",
                    "runtime" : 18000 },
                "1" :
                {  
                    "name" : "simple_acces",
                    "desc" :  "access items at 1k ops with 80% gets",
                    "workload" : "s:15,g:80,d:5,ccq:simplekeys,ops:4000",
                    "template" : "default",
                    "runtime" : 18000 }
        }
}

ops/sec is now at 20k
i also started on mcsoda : root@ubu-1812:~# lib/perf_engines/mcsoda.py membase://10.3.2.105:8091 max-items=1000000 json=1 threads=2 vbuckets=1024 ratio-sets=0.3 doc^Cen=0 doc-cache=0


system test workers from 10.3.121.44 and 10.3.121.46 ,mcsoda is from 10.3.121.47

i am not running stats collector yet

now i think i added a worker that should collect stats because 
WORKERS = ['10.3.121.47']
# valid configs ["kv","query","admin","stats"] or ["all"]
WORKER_CONFIGS = ["stats"]
CB_CLUSTER_TAG = "plum"
ATOP_LOG_FILE = "/tmp/atop-node.log"
LOGDIR="logs"  # relative to current dir

are you going to use this cluster for upgrading to 1971 ? or is this plum cluster which we want to keep it running for many more day s?

ok this is plum cluster - so i assume we want to keep this running for a while

• copying from skype :
  [12/2/12 1:45:49 PM] Farshid Ghods: and regarding plum , since it does not have chiyoung fix for memory leak in warmup
  [12/2/12 1:46:51 PM] Farshid Ghods: please verify that you SEE memory leak