As was mentioned in HA Deployment, an SSL load balancer is required, which will decrypt SSL connections and also balance requests across the Helix TeamHub Web servers.
The Helix TeamHub package does not include load balancer, therefore it needs to be installed separately. The following guide is recommended to setup the load balancer first if none exists.
With a load balancer, the user requests will be randomly distributed across the cluster nodes, so the data will become immediately out of sync. To fix this issue, attach the same shared storage to each Helix TeamHub Web server. If existing storage with a clustered file system doesn't exist, contact the Support team for further help.
After shared storage is available, stop Helix TeamHub, then mount the storage to
/var/opt/hth/shared. Next, bring Helix TeamHub back online:
sudo hth-ctl stop sudo mv /var/opt/hth/shared /var/opt/hth/shared.bak # Mount storage to /var/opt/hth/shared and sync back the data sudo rsync -av /var/opt/hth/shared.bak/ /var/opt/hth/shared/ rm -rf /var/opt/hth/shared.bak sudo hth-ctl start
Since the SSH host keys will differ between the cluster nodes, they need to be synchronized. Helix TeamHub configuration process can use the
ssh directory on the shared storage and copy the SSH host keys to the usual
/etc/ssh. This will enable every new Helix TeamHub Web server added to the cluster to have the same SSH host keys. So on the first Helix TeamHub Web server, create the directory and copy SSH host keys:
mkdir -p /var/opt/hth/shared/ssh cp /etc/ssh/ssh_host_* /var/opt/hth/shared/ssh/ chown root.root /var/opt/hth/shared/ssh/* chmod 600 /var/opt/hth/shared/ssh/*
After steps 1 through 3 are performed, additional Helix TeamHub Web servers can be added to the cluster. Because the Helix TeamHub Configuration file is stored on a shared partition
/var/opt/hth/shared, simply install the Helix TeamHub Web package and reconfigure it, and Helix TeamHub will automatically pick up the needed configurations:
yum install python-devel gettext python-docutils pam-devel cronie postfix libidn libidn-devel sudo bsip2 mailx libgomp rpm -ivh hth-web-X.X.X-stable.el6.x86_64.rpm
apt-get install python-dev python-docutils gettext sudo libpam0g-dev postfix libidn11 libidn11-dev cron bzip2 mailutils libgomp1 dpkg -i hth-web_X.X.X_amd64.deb
Internet Site during setup.
At this stage everything should be up and running, and requests should be distributed across all Helix TeamHub Web servers. However, Helix TeamHub is still bootstrapped with the hostname of the first Helix TeamHub Web server installed. To fix this, go to Helix TeamHub Admin and change the hostname to the load balancer.