Frequently Asked Technical Questions

A sudden spike in CPU usage on Linux servers is one of the most common causes of performance degradation. The following tools should be used to analyze this situation:

• top and htop: Display real-time CPU usage
• ps aux --sort=-%cpu: Lists the processes using the most CPU
• uptime: Provides load average values

In particular, high load average values (e.g., exceeding the number of CPU cores) indicate that the system has reached a bottleneck.

You should install updates regularly to patch security vulnerabilities and improve performance:

• Ubuntu/Debian: sudo apt update && sudo apt upgrade -y
• CentOS/AlmaLinux/Rocky Linux: dnf update -y or yum update -y
• Windows: You should install the latest "Security Patch" and "Quality" updates via the "Check for Updates" menu and then restart the server.

Swap refers to the virtual memory space created on the disk to prevent the server from freezing when RAM usage is limited. In Linux, you can create it by following these steps:

• fallocate -l 2G /swapfile (allocates 2 GB of space)
• chmod 600 /swapfile (Sets permissions)
• mkswap /swapfile and swapon /swapfile (Enables the swap space)
• Don’t forget to add it to the /etc/fstab file to make it permanent.

If you prefer to manage your server using a graphical interface instead of the command line, you can install popular control panels. Most panels are installed on a clean operating system (Clean OS) using a single-line script. For example, to install CyberPanel, you can enter the command sh <(curl https://cyberpanel.net/install.sh || wget -O - https://cyberpanel.net/install.sh) in the terminal to launch the installation wizard.

NOTE: Before installation, make sure your server meets the minimum RAM requirements.

Port 22 is the port most frequently targeted by brute-force attacks. To improve security:

1. Open the /etc/ssh/sshd_config file with a text editor.
2. Remove the comment from the #Port 22 line and update it to Port 2222 (or an unused port).
3. Important: Before saving the changes, allow the new port through the firewall (UFW/Firewalld).
4. Restart the service using the systemctl restart sshd command.

When you increase your disk capacity or add a new disk via the customer panel, the operating system must be configured at the system level for this space to become available:

• Linux: Use fdisk -l to locate the new disk, format it with mkfs.ext4, and mount it to a directory, then add it to the /etc/fstab file.
• Windows: Open the "Disk Management" tool, initialize the new disk, and create a new simple volume to assign a drive letter.

Anycast DNS involves hosting your domain’s DNS records simultaneously on multiple servers located in various geographic regions around the world. When a user attempts to access your site, the DNS query is routed to the server physically closest to them. This technology reduces DNS resolution time to the millisecond level and enhances your site’s availability by distributing the load to mitigate potential DDoS attacks.

If you have multiple servers (for example, an application server and a database server), you can configure these servers to communicate with each other over a local network (VLAN) without accessing the internet. This configuration:

• Security: Ensures that database traffic remains isolated from the outside world.
• Speed: Minimizes latency by leveraging local network speeds of 1 Gbps or 10 Gbps.
• Cost: Data traffic passing through the local network is generally not deducted from your bandwidth quota.

• A Record: Directs a domain name to an IPv4 address (e.g., 185.x.x.x). It is the fastest type of record.
• CNAME Record: Redirects a domain name to another domain name (alias) (e.g., blog.site.com -> site.com).
  Important Rule: Due to technical standards, a CNAME record cannot be used for the root domain (site.com); an A record must be used instead. For subdomains, however, a CNAME record is preferred to allow for greater flexibility.

A PTR record is a "reverse" record that indicates which domain name an IP address is associated with. It is particularly critical for users who manage their own mail servers. Most email service providers (Gmail, Outlook, etc.) check the sender’s IP address PTR record to verify the legitimacy of incoming emails. If your PTR record does not match the domain you’re sending from, your emails may end up in the "Spam" folder or be rejected.

When you change a DNS record (A, CNAME, MX, etc.), it typically takes between 2 and 24 hours for the change to propagate to all DNS servers worldwide. This is related to the TTL (Time to Live) value associated with the records. Until the TTL expires, internet service providers keep the old records in their cache. To speed up the process, you can lower the TTL value before making the change.

Traditional passwords can be compromised through brute-force attacks. With the SSH key system, however, you connect to the server using a "private key" file that you physically possess.

• Advantage: It becomes impossible to gain access to the server through trial-and-error password attempts.
• Procedure: After adding the "Public Key" you created with ssh-keygen to the authorized_keys file on the server, you can ensure the highest level of security by completely disabling password-based login (set PasswordAuthentication to no).

Your server management panel (Netinternet Customer Panel) is the key to your entire infrastructure. Enabling 2FA (Two-Factor Authentication) via Google Authenticator or SMS when logging into the panel prevents your servers from being deleted or your data from being stolen, even if your password is compromised. This acts as a "second lock" for your digital assets.

An SSL certificate prevents "man-in-the-middle" attacks by encrypting the data transmitted between the user and the server. However, today it is not only a security requirement but is also essential for the following reasons:

• SEO: Google pushes websites without an SSL certificate to lower rankings in search results.
• Trust: The "Not Secure" warning in browsers leads to user loss.
• Payment Systems: The use of SSL (TLS 1.2+) is a legal requirement for accepting credit card payments.

Fail2ban is a security tool that continuously scans server logs. It automatically blocks IP addresses that exceed the specified number of failed login attempts by using the system’s firewall (iptables). It is a critical tool for stopping bot attacks targeting SSH, FTP, and SMTP services.

Your server's security starts with opening only the necessary ports to the outside world. By default, you should close all ports and allow access only for the services you use:

• Linux (UFW): You can allow web traffic using the commands ufw allow 80/tcp (HTTP) and ufw allow 443/tcp (HTTPS).
• Windows: You can set restrictions in the "Inbound Rules" section of the "Windows Firewall with Advanced Security."
• Recommendation: Opening SSH or RDP ports exclusively for your own static IP address will reduce your attack surface by 99%.

Storing backups on the same disk or in the same physical cabinet as the server is a major risk. In the event of a hardware failure, file system corruption, or ransomware attack, both your primary data and your backups could be destroyed simultaneously. Off-site backup is the "last refuge" for your data during a disaster.

This golden rule is the foundation of security: You should have at least 3 copies of your data, stored on 2 different storage technologies (e.g., disk and cloud), and 1 copy must be kept in a physically different location (Off-site) from the data center where the server is located.

Snapshot is a quick copy of your server's state (including file system and RAM) at a specific moment, usually used as a "restore point" before updates. Backup, however, is the copying of your data to an independent storage unit outside the server. Snapshots are not suitable for long-term storage; regular backups are essential for true data security.

To find performance issues at their source, you should enable the "Slow Query Log" feature. This logs queries that take longer than a specified duration (e.g., 1 second) to a file. You can then analyze these queries using the EXPLAIN command to easily see where bottlenecks occur and identify missing indexes.

Indexes allow the database to search through an "index" rather than scanning the entire table (Full Table Scan). Properly applied indexes to specific columns can speed up SELECT queries thousands of times. However, keep in mind that every unnecessary index slows down INSERT and UPDATE operations because the index must also be updated with each transaction.

The fastest method for taking a backup via the command line is the mysqldump tool. Use mysqldump -u user -p db_name > backup.sql to export and mysql -u user -p db_name < backup.sql to restore. For large databases, performing this action in a screen or tmux session protects you against connection drops.

This error indicates that the number of clients trying to connect simultaneously exceeds the max_connections limit. To resolve it, you can increase this limit in the database configuration file or optimize your application's "Connection Pool" settings to ensure connections are properly closed once their task is complete.

For security reasons, databases are open only to the local network (localhost) by default. To allow remote access, you must change the bind-address value to 0.0.0.0 in the my.cnf file for MySQL/MariaDB and grant privileges to the user for your specific IP address. Also, ensure that port 3306 (MySQL) or 5432 (PostgreSQL) is open in the firewall.