mirror of
https://github.com/corda/corda.git
synced 2024-12-26 16:11:12 +00:00
83fdf678ab
* Update doc to show JMX no longer works in Corda, due to the serialisation changes. * minor changes * address PR issues
80 lines
4.9 KiB
ReStructuredText
80 lines
4.9 KiB
ReStructuredText
Node administration
|
|
===================
|
|
|
|
When a node is running, it exposes an RPC interface that lets you monitor it,
|
|
you can upload and download attachments, access a REST API and so on. A bundled
|
|
Jetty web server exposes the same interface over HTTP.
|
|
|
|
Logging
|
|
-------
|
|
|
|
By default the node log files are stored to the ``logs`` subdirectory of the working directory and are rotated from time
|
|
to time. You can have logging printed to the console as well by passing the ``--log-to-console`` command line flag.
|
|
The default logging level is ``INFO`` which can be adjusted by the ``--logging-level`` command line argument. For more
|
|
custom logging, the logger settings can be completely overridden with a `Log4j 2 <https://logging.apache.org/log4j/2.x>`_
|
|
configuration file assigned to the ``log4j.configurationFile`` system property.
|
|
|
|
Database access
|
|
---------------
|
|
|
|
The node exposes its internal database over a socket which can be browsed using any tool that can use JDBC drivers.
|
|
The JDBC URL is printed during node startup to the log and will typically look like this:
|
|
|
|
``jdbc:h2:tcp://192.168.0.31:31339/node``
|
|
|
|
The username and password can be altered in the :doc:`corda-configuration-file` but default to username "sa" and a blank
|
|
password.
|
|
|
|
Any database browsing tool that supports JDBC can be used, but if you have IntelliJ Ultimate edition then there is
|
|
a tool integrated with your IDE. Just open the database window and add an H2 data source with the above details.
|
|
You will now be able to browse the tables and row data within them.
|
|
|
|
Monitoring your node
|
|
--------------------
|
|
|
|
Like most Java servers, the node exports various useful metrics and management operations via the industry-standard
|
|
`JMX infrastructure <https://en.wikipedia.org/wiki/Java_Management_Extensions>`_. JMX is a standard API
|
|
for registering so-called *MBeans* ... objects whose properties and methods are intended for server management. It does
|
|
not require any particular network protocol for export. So this data can be exported from the node in various ways:
|
|
some monitoring systems provide a "Java Agent", which is essentially a JVM plugin that finds all the MBeans and sends
|
|
them out to a statistics collector over the network. For those systems, follow the instructions provided by the vendor.
|
|
|
|
.. warning:: As of Corda M11, Java serialisation in the Corda node has been restricted, meaning MBeans access via the JMX
|
|
port will no longer work. Please use java agents instead, you can find details on how to use Jolokia JVM
|
|
agent `here <https://jolokia.org/agent/jvm.html>`_.
|
|
|
|
`Jolokia <https://jolokia.org/>`_ allows you to access the raw data and operations without connecting to the JMX port
|
|
directly. The nodes export the data over HTTP on the ``/jolokia`` HTTP endpoint, Jolokia defines the JSON and REST
|
|
formats for accessing MBeans, and provides client libraries to work with that protocol as well.
|
|
|
|
Here are a few ways to build dashboards and extract monitoring data for a node:
|
|
|
|
* `JMX2Graphite <https://github.com/logzio/jmx2graphite>`_ is a tool that can be pointed to /monitoring/json and will
|
|
scrape the statistics found there, then insert them into the Graphite monitoring tool on a regular basis. It runs
|
|
in Docker and can be started with a single command.
|
|
* `JMXTrans <https://github.com/jmxtrans/jmxtrans>`_ is another tool for Graphite, this time, it's got its own agent
|
|
(JVM plugin) which reads a custom config file and exports only the named data. It's more configurable than
|
|
JMX2Graphite and doesn't require a separate process, as the JVM will write directly to Graphite.
|
|
* Cloud metrics services like New Relic also understand JMX, typically, by providing their own agent that uploads the
|
|
data to their service on a regular schedule.
|
|
* `Telegraf <https://github.com/influxdata/telegraf>`_ is a tool to collect, process, aggregate, and write metrics.
|
|
It can bridge any data input to any output using their plugin system, for example, Telegraf can
|
|
be configured to collect data from Jolokia and write to DataDog web api.
|
|
|
|
Memory usage and tuning
|
|
-----------------------
|
|
|
|
All garbage collected programs can run faster if you give them more memory, as they need to collect less
|
|
frequently. As a default JVM will happily consume all the memory on your system if you let it, Corda is
|
|
configured with a relatively small 200mb Java heap by default. When other overheads are added, this yields
|
|
a total memory usage of about 500mb for a node (the overheads come from things like compiled code, metadata,
|
|
off-heap buffers, thread stacks, etc).
|
|
|
|
If you want to make your node go faster and profiling suggests excessive GC overhead is the cause, or if your
|
|
node is running out of memory, you can give it more by running the node like this:
|
|
|
|
``java -Xmx1024m -jar corda.jar``
|
|
|
|
The example command above would give a 1 gigabyte Java heap.
|
|
|
|
.. note:: Unfortunately the JVM does not let you limit the total memory usage of Java program, just the heap size. |