I would like to create VPS infrastructure like some from these variants Cloud Computing System Architecture Diagrams
I have some questions?
- Which variant you using?
- What about when IW master cluster node will shut down?
- Do you recommend another VPS infrastructure?
Our clustering solution is not HA. The cluster manager is always a single point of failure–if it goes down, everything will.
Hi Standus, Hi All.
Since Interworx does not provide an HA solution, we solved it with Virtualization.
We have installed the cloud-virtualization platform XCP-NG (formerly xenserver) on our servers.
It is necessary to use at least three servers and a shared iSCSI storage, where the disks of the virtual servers are physically stored.
Servers are organized into POOLs.
Then we created multiple VMs on the servers and Interworx control panel is installed in each one.
1pcs Cluster Master node
4pcs Cluster Work nodes
- Cluster Master node can be incrementally backed up to some other HW server in another POOL (due to Disaster recovery - DR), every 3600 seconds.
- Cluster Work Nodes are divided by servers and this ensures HA for Cluster nodes = WorkNodes. These can be incrementally backed up to some other HW server in another POOL (due to Disaster recoveryDR), every 3600 seconds.
So how does HA work?
The virtualization platform itself will ensure that in the event of a HW server failure, it will automatically start a VM from another HW server. This applies equally to CLuster Master and Cluster Nodes.
Usually, the customer does not even notice that something has happened in the virtualization and web hosting infrastructure.
In case something more serious happens, we always have DR VM copies available in another POOL of servers, and in this case the recovery time is equal to the DR VM copies startup time.
Of course, DR copies can also be created in another location, this is also considered in the virtualization technology.
Such a solution has worked for us…