Benoit Perroud created CASSANDRA-11821:
------------------------------------------
Summary: nodetool cfhistograms throws
java.lang.ArrayIndexOutOfBoundsException and reset the cfhistograms
Key: CASSANDRA-11821
URL: https://issues.apache.org/jira/browse/CASSANDRA-11821
Project: Cassandra
Issue Type: Bug
Components: Tools
Environment: 2.0.17, rhel6.4
Reporter: Benoit Perroud
Priority: Minor
I had nodes running for a decent amount of time (like 3 months).
{code}
nodetool cfhistograms keyspace table
{code}
did throw and {{ArrayIndexOutOfBoundsException}} in the read latency section:
{code}
Read Latency (microseconds)
Exception in thread "main" java.lang.ArrayIndexOutOfBoundsException: 90
at org.apache.cassandra.tools.NodeCmd.printHistogram(NodeCmd.java:1086)
at
org.apache.cassandra.tools.NodeCmd.printCfHistograms(NodeCmd.java:1138)
at org.apache.cassandra.tools.NodeCmd.main(NodeCmd.java:1472)
{code}
Surprisingly, the running the same cmd a second time did succeeded, but the
histograms seem to be reset:
first call:
{code}
SSTables per Read
1 sstables: 183026
2 sstables: 176538
3 sstables: 49159
4 sstables: 23034
5 sstables: 16380
...
Write Latency (microseconds)
3 us: 4
4 us: 33
5 us: 237
6 us: 1193
7 us: 4055
8 us: 14357
10 us: 146728
12 us: 565588
14 us: 1907616
17 us: 8468503
20 us: 16218034
...
{code}
After:
{code}
SSTables per Read
2 sstables: 52
3 sstables: 4
Write Latency (microseconds)
6 us: 1
7 us: 3
8 us: 5
10 us: 53
12 us: 133
14 us: 379
17 us: 1190
20 us: 1605
{code}
I didn't find anything on the logs, and I wasn't able to have the exception
popping out more than once per servers.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
