YorkHost - Notice history

The issue has been fully resolved and the Client Area is now functioning normally.
We apologize for the disruption experienced. The recent deployment of our new high-availability HAProxy setup caused an unexpected configuration issue, which temporarily affected access through Cloudflare.

While high availability improves resilience, it can occasionally create unexpected behavior during synchronization. Our team has fine-tuned the configuration to avoid future occurrences and continues to monitor performance closely.

We are aware that the Client Area may still experience slow response times (around 3–4 seconds per refresh).
We continue to work on optimizing this, but debugging remains complex since we are not the developers or owners of the CMS (HostBill).
As a result, our main efforts are focused on the new Client Area Beta, which, being built entirely in Node.js, is far more stable, resilient, and easier to maintain.

Thank you for your continued patience and trust.

The degraded performance detected earlier was caused by an incorrect HAProxy configuration following last night’s infrastructure enhancement.
After recovering backups, we introduced multi-replica synchronization between the master and several secondary nodes.
An error in the configuration caused the HAProxy service to stop, leading Cloudflare to detect the backend as unavailable.

The issue has now been corrected, and all systems are operational.
We continue to monitor performance to ensure stability.

The root cause was identified as a link flap protection trigger (linkFlapErrDisabled) on interface Ethernet104/1/47 of edge1-par3-b9.
The port has been manually re-enabled after confirming stability, and network access to PVE-03 has been restored.

The node is now fully operational, and services are back online.
Monitoring will continue to ensure stability following the recovery.

The system has been repaired in rescue mode, and we are now attempting to restart and access the internal Proxmox environment.
Initial recovery steps completed successfully, and network reinitialization is underway.

Monitoring will continue throughout the reboot process.
Further updates will follow once the node status is confirmed stable.

Our team is currently working in rescue mode using a Debian image to repair the affected Proxmox environment.
The issue originated from a memory allocation failure, which also disrupted the network configuration.

Restoration efforts are in progress to bring the node back online safely.
We will share another update once the system is stabilized and network access is restored.

The node PVE-03 experienced a crash due to a memory allocation error.
As a result, the system is currently facing network recovery issues.

Our technical team is actively investigating and working to restore normal connectivity as quickly as possible.
Further updates will be provided as soon as progress is made.

We appreciate your patience and understanding.

After reviewing multiple tickets, the resolution has been confirmed.
Please update your FiveM artifacts to version:
21547-0b6d5de3902cda6dd91be0c489d9b7243e554bb1
All affected instances are now fully operational. Monitoring continues to ensure stability across all nodes.

Update:
The issue has been resolved after updating the FXServer artifacts to version 21547-0b6d5de3902cda6dd91be0c489d9b7243e554bb1.
All affected instances are now operating normally, and no further crashes have been observed.

Resolution details:

• Updated FXServer artifacts to the latest stable build

• Confirmed compatibility with txAdmin and the new TXHOST_* environment variables

• Deprecated ConVars removed from all configurations

Monitoring will continue for next hours to ensure full stability across all nodes.

HAPROXY & Beta Client nodejs build is fixed now.
It should be come back in next minutes.

Client Area restored with temporaly web server, Beta Client still under maintenance.

The geoblock remains active to limit the impact.

A mitigation patch is still being tested with our service provider.

Some users may experience slowdowns or temporary unavailability.

A modification is being made to mitigate the attack.

We have implemented geographic restrictions while waiting for the Anti-DDoS provider’s patch, scheduled for tomorrow.

Protecting Plesk servers remains challenging due to their shared infrastructure, but we are actively working on long-term improvements.

The service is currently partially restored.

Le service Plesk est toujours visé par une attaque de niveau L7.
Nos équipes poursuivent le déploiement d’un correctif pour rétablir la stabilité.