Urgent SolarWinds migration of large number of nodes

Successfully implemented thanks to knowledge, insight and with the help of a self-developed provisioning tool.

This SolarWinds migration between two merging international service providers was extremely difficult. Procyon Networks has therefore been urgently asked to migrate more than 5,000 nodes from one SolarWinds installation to another. These 5000+ nodes represent as many customers with associated service contracts and are therefore our customer's core business. The SolarWinds migration had to be carried out quickly and error-free to ensure that the services continued to run smoothly.

Problems during the migration
During the first attempt the following problems arose:
It was decided to perform the migration using an exported list of nodes. A scan was made from the new system with this list. The result was that nodes that were inaccessible from the new system were not registered.
The new nodes were only known by their IP addresses. All information regarding naming and custom properties had to be added afterwards. The adjustment of IP addresses to correct (informative) names went wrong for a large number of nodes. Subsequently, many custom properties were assigned to the wrong nodes.

By attempting to correct the errors in the process by rescanning and importing, even nodes were duplicated.

After several attempts to solve the aforementioned problems, the migrated data turned out to be completely corrupt and Procyon Networks was called to save the day and the migration.

How was Procyon Networks able to successfully complete the migration so quickly?
We had already developed a preliminary provisioning tool for the customer who uses the source system. With this tool, data from a CMS could be used to create the correct nodes in SolarWinds. After some adjustments, our tool was able to import all nodes directly and correctly with a full export from the CMS via the SolarWinds API. In addition, we have created some ad-hoc scripts especially for recognizing duplicate and corrupt data and for reporting. This allowed Procyon Networks to show exactly why and when the errors occurred.

The final migration took barely 2 hours for the 5000+ nodes, which allowed us to help our customer within their intended deadline.

Large-scale SolarWinds migration? Keep it simple
Working with multiple lists, the data of which must be manipulated and merged, is very error-prone. This has proven once again in this situation. In addition, rescanning large numbers of nodes is not a reliable method. In the new situation, not all nodes need to be immediately accessible (think of downtime and firewall issues). However, unreachable nodes must also become available in the new system.

For these format migrations, adding all information correctly and completely is of course very important. For this, working with a single data source and using the SolarWinds API is by far the best method. In addition, the speed of such a migration is important for a provider because changes cannot be implemented during this migration.

The challenge is bigger than the solution. With knowledge and insight it is possible to carry out an extensive operation quickly, easily and reliably.