-
Notifications
You must be signed in to change notification settings - Fork 1.2k
Installation on Cloudera Virtual Machine
This is a quick guide for installing on Clouderas pre-built Hadoop-machines.
First, download and install a Cloudera Demo VM (I fetched a copy of Download installer
The machine will need at least 3 GB of RAM in order to compile OpenTSDB while running the Hadoop-stuff. If you kill all things Java before compiling (sudo pkill java, then compile and finally reboot to get everything back up), you can get by with 2 GB.
In the likely case you want to send data to the machine, you will need to set it up with Bridged Networking (rather than NAT, which is the default many places).
Once you got the machine booted, you might notice that it is set up with Eastern Standard Time (EST). (I didn't. Twice.) In order to change it to UTC (or whatever else), run the following:
sudo ln -sf /usr/share/zoneinfo/UTC /etc/localtime
sudo yum install git automake gnuplot
git clone https://github.com/OpenTSDB/opentsdb.git
cd opentsdb
./build.sh
env COMPRESSION=none HBASE_HOME=/usr/lib/hbase ./src/create_table.sh
mkdir /tmp/tsd
Finally, start TSD with:
./build/tsdb tsd --port=4242 --staticroot=build/staticroot/ --cachedir=/tmp/tsd/
You should now be able to play with OpenTSDB on http://<virtual machine ip>:4242.