Collector: Install to an additional cluster node

If the configuration of  reload and its collector on the first node is done and successful tested, you want to install the GWDR-Collector at the other nodes. The documentation to do this is ...

I have archived this by a few simple lines at bash from each node where the collector should added.

Get the software and install.
# cd /root/inst-sys
# scp <reload-server>:/opt/beginfinite/reload/setup/collect/gwdr-collector.rpm
# zypper in gwdr-collector.rpm

(You get and hint the the package is not signed. Ignore this. Maybe some guy from MF will correct this)

Generate and distribute SSH-Key:
# ssh-keygen -t rsa
# ssh-copy-id -i /root/.ssh/id_rsa.pub <reload-server>

Check connection (optional):
# ssh <reload-server>

Copy configuration:
# cd /opt/beginfinite/reload/conf/
# scp -r <1st-node>:/opt/beginfinite/reload/conf/* .
# scp <1st-node>:/opt/beginfinite/reload/license/reload.pem ../license/

Search specific configuration of the 1st-node:
# find -type f -exec grep -l <Seachstring> {} \;

Seachstring could be the IP-Address or the Hostname.

Tune configuration:
# find -type f -exec sed -i "s/<old>/<new>/g" {} \;

Example:
# find -type f -exec sed -i "s/10.0.0.1/10.0.0.2/g" {} \; or
# find -type f -exec sed -i "s/1st-node/2nd-node/g" {} \;

(optional, Test)
# find -type f -exec grep -l <Searchstring> {} \;

Start the collector daemon:
# reloadd restart

Test if Web-UI is running:
# ss -tulpen | grep :5555

==========================

During the installation of the 1st node I used the hostname and IP-address of the GW-resource. (In this case there is only one ressource for all GW-PO and -DOM.)

If you have more than one GW-Resource it cloud be possible to run the GWDR-Daemon one time per resource. Check # /opt/beginfinite/reload/lib/code/gre --help for the cmd line parameters.

Tags:

  •  don´t understand what you mean with this :

    If you have more than one GW-Resource it cloud be possible to run the GWDR-Daemon one time per resource. Check # /opt/beginfinite/reload/lib/code/gre --help for the cmd line parameters.

    Normally is GWDR-Daemon running multiple times per GW-Resource ?

     

    Thanks

     

    P. S. How big is your cluster ? Number of nodes and numbers of Post Offices ...

  • :
    My post was nearly one year old
    I thought ... ahh ... check the help page. You can start more than one reload process at a time. So if your LAN is perfomant you can save backup time.

    There are some cluster enviroments from 2 to 4 nodes. And they run GW from 1 to 10 PO (~100GB - 4TB)