Real case: how a construction company lost 730 days on software installation "due to a socket"

How it all started

The director of the production branch of the construction company is tired of tolerating the chaos in internal processes. The company has factories, warehouses, and a central office, but there is no unified document management system. The accounting department works in 1C, tasks are lost in emails, and reports are stored in Excel on different computers.

With several projects being launched simultaneously, control disappears, deadlines are missed, responsible parties cannot find the necessary documents, and information scatters across folders and correspondence. The director decides to implement a document management system. The management purchases a local version of the Directum Lite system and assigns the IT specialist the task: to implement the DMS with a deadline of "yesterday."

First mistake: they do not consult with the IT specialist regarding the choice of software and the readiness of the infrastructure for the installation of the system in the company.

The implementation of Directum began with the installation of software on the server

Data must be stored within the corporate perimeter. The company decides to prepare the infrastructure independently, without help from the vendor. The responsible person has no experience working with such tasks.

First failure

Three days after the implementation of the system, the server burns out. We could not determine why this happened. But the work on launching the DMS had to be postponed until new equipment was delivered.

Two months later, a new server is delivered, and after another three months, some data is managed to be restored. In total, five months pass.

To speed up the software installation, the construction company orders system configuration from the vendor's specialists. And finally, they install the DMS on the server in just a few days.

Second failure

The system is set up; it's time to launch the processes. But after a while, the electricity goes out in the building. A voltage spike damages the SSD drives on the server. The data is inaccessible, and work stops even before it has begun.

Error two: The company is restoring the system, but they are not drawing conclusions about the infrastructure. They are not purchasing a backup power supply, not increasing the number of servers, not distributing them across different offices, and not ensuring fault tolerance.

Third failure

The system is up and running, but history repeats itself. Another voltage spike and the SSD drives fail again. The IT specialist manually re-deploys the system: installation, configuration, data input.

Conclusion

Due to the lack of measures taken to protect against power interruptions and establish server infrastructure, the company spends two years installing software. 730 days are spent on finding problems and causes of failures, recovering lost data, cyclically installing the same system, and studying technical documentation.

After six months, the customer stops responding. The organization has not managed to quickly establish document flow.

How to avoid repeating mistakes

“The first rule of Fight Club”: check the infrastructure before launching the system.

Protect server equipment from voltage spikes

Install an uninterruptible power supply and configure proper system shutdown in case of power loss. This will mitigate voltage fluctuations and extend the life of the equipment.

Set up backups

Regularly back up the database and the file storage of the system's configuration files. In case of a failure, you will restore the system from a copy instead of setting it up from scratch. This is cheaper and faster than reinstalling and populating the system with data.

Eliminate single points of failure

In a production environment, there should be no single-instance nodes in the system. An additional server will improve fault tolerance. If one fails, the other will keep operations running.

Use reliable equipment

Purchase specialized server hardware. For example, storage devices and quality power supplies. During a voltage spike, the power supply absorbs the impact, protecting the drives.

Saving on hardware will lead to costs for replacing equipment. This will take time and provoke risks of business downtime: operations will halt, and losses will increase.

Check the power supply line

Make sure that the server hardware is connected to a stable power line. If necessary, install stabilizers to minimize the strength of surges. This way, the equipment will not be harmed due to problems with the building's power grid.

Instead of output

The infrastructure is invisible until it breaks down. But when that happens — the business stops. The implementation of software starts not with purchasing a license, but with preparing the "hardware." Such security is created by qualified engineers. Invest in infrastructure and training specialists before starting the implementation of the document management system.