Phoronix.com faces brief downtime for NVMe RAID drive replacement

Phoronix.com and its associated sites will experience brief downtime around midnight PST / 3 AM EST on November 1, 2026, due to a failed drive in the site's NVMe SSD RAID array. The hardware cannot be hot-swapped, requiring hosting provider HiVelocity to power down the server for physical replacement.

The failed drive was replaced just one year ago. Site operator Michael Larabel attributes the accelerated hardware wear to increased traffic from AI scraping bots and similar automated traffic that has intensified throughout 2026.

Infrastructure and hardware details

• RAID array type: The site runs on an NVMe SSD RAID configuration, indicating redundant solid-state storage for performance and data protection.
• Hot-swap limitation: The failed drive cannot be replaced without powering down the server, suggesting either a hardware backplane limitation or a specific RAID controller implementation that lacks online drive replacement capability.
• Replacement frequency: The one-year replacement interval for a failed enterprise NVMe drive points to unusually heavy sustained I/O load on the storage subsystem.
• Traffic attribution: AI scraping bots are identified as a significant contributor to site load, reflecting broader infrastructure pressure on technical publications from automated data collection systems.

Why it matters

Phoronix serves as a primary benchmark and Linux hardware news source for the open-source community, making its availability relevant to developers tracking driver performance and kernel developments. The incident also illustrates how AI scraping traffic is creating tangible infrastructure costs and hardware wear for independent technical publishers, a growing concern for sites that provide free access to benchmark data and open-source coverage.