Can you provide the output from `nodetool tpstats`. -Chris
On Jul 20, 2010, at 8:59 PM, Dathan Pattishall wrote: > Type 'help' or '?' for help. Type 'quit' or 'exit' to quit. > cassandra> connect cass01/9160 > cassandra> get TimeFrameClicks.Standard2['test_cassandra_alive'] > Exception null > > The data exists and I can grab the data after I restart all the nodes, > but once the cluster runs for a few minutes I cannot grab this > specific key or random other keys. It takes about 3 seconds until the > Exception null message. My storage-conf.xml is very simple: > > <Keyspace Name="TimeFrameClicks"> > <ColumnFamily Name="Standard2" CompareWith="UTF8Type" /> > .... > > Now my data is very small like 20 GB across 4 servers. Writes > consistently remain fast, yet reads fail like crazy. I hope its > something that I am doing wrong because > > nodetool cfstats > > says that the read latency for the keyspace and this specific column > family is less then 0.3 ms which means that something is lying to me. > > To head off some questions: > > CPU utilization is very little. > There is hardly any I/O on the box > The servers are all the same class enterprise boxes > There is 12 GB of ram per server > Each Server uses a local RAID. > Nothing in any of the system logs that indicates there any problem. > > Additionally is there a stat or series of stats that I can lookup to > determine the health of read performance.
