Jolokia version mismatch results in very large exception message.

[MaxFirmino]

Description:

1- Does the problem persist?
Yes.

2- Describe how to reproduce the issue, if possible. If not, describe why.
Attempt to connect a Cassandra node to ecChronos Agent using a Jolokia version prior to 2.5.1. Eventually, this will result in a known IllegalStateException, with a message that starts roughly like so:

Unexpected error during schedule change of keyspace_name.table_name: javax.management.RuntimeMBeanException: java.lang.IllegalStateException : Internal: More than one key found when extracting with path:<very very very long string of paths>

Since this is a known problem, there is a known solution, which is to simply use the correct version of Jolokia. This however is still a problem for upgrade scenarios, where a service using the original ecChronos upgrades to ecChronos Agent.

Detailed description:
• What is happening?
The exception message is not very clear, is too big and happens for every table of every keyspace, which can cause confusion.

• What did you expect to happen?
The exception should be treated and given a clearer message. Perhaps the original, full exception message can still be logged at DEBUG level.

• What version of ecChronos are you using?
Master of ecChronos Agent

• Was the problem detected during an upgrade or downgrade procedure?
Yes.

• Have you checked the ecChronos documents?
Yes.

Logs available upon request.

[tommystendahl]

The old ecChronos did't use Jolokia.
You can use Jolokia Agent versions prior to 2.5.1 with the Jolokia 2.5.1 used in ecChronos. Jokolia Agents 2.4.x is tested and works fine with ecChronos with Jokolia 2.5.1.

[tommystendahl]

The valid scenario for this issue is if you are are already using Jolokia before installing ecChronos-agent and you are using a jolokia agent with version 2.3.0 or older and don't upgrade the jolokia agents before installing ecChronos-agent.