The developers of the systemd project have identified and are now addressing a significant architectural issue in the “systemd-journald” component. This issue has led to a substantial overestimation of the data written to disk, known as write amplification, compared to the actual volume of logs.
The origins of this problem trace back to March 2020 when a report demonstrated that generating approximately 500 KB of text logs resulted in over 700 MB of physical writes to the SSD. Initially, the systemd developers refuted the existence of the problem, dismissing it as a misunderstanding of file systems. Despite receiving backlash from users, the project’s stance remained unchanged.
In early 2026, independent developer ValdikSS submitted a detailed report highlighting the issue. Through meticulous profiling using isolated cgroups and loop devices, he illustrated how memory-mapped files and binary hash tables caused significant block device-level I/O when writing minimal text messages. This report shed light on the root cause of the problem.
Following recent attention on Hacker News and compelling synthetic tests, the project maintainers have revamped their approach. They are now focusing on optimizing cache reset mechanisms and restructuring the journald index storage to mitigate the write amplification issue. This shift in strategy reflects a corporate Open Source mentality, where a critical flaw is addressed only when the damage to the project’s reputation outweighs the cost of fixing it.