Hallo allerseits
In meinem Ticket habe gestern nun dieses Feedback vom Support erhalten:
--------------------------------------------------------------------------------------------------------
Thank you very much for sharing your experience with us.
From what you described, the performance difference seems to appear mainly after enabling SSD caching. Your sequential transfer rates already saturate the 2.5 GbE network, which suggests that the network, disks, and baseline throughput are all working properly. This makes the caching layer itself the most likely factor behind the performance gap.
Currently, our NAS uses dm-cache, which is generally strong for large sequential data transfers but is less efficient for handling random I/O workloads. This often results in noticeable differences compared to systems that use more optimized caching mechanisms.
This is also why our engineering team has placed a high priority on evaluating bcache. Bcache is designed to handle random access patterns more effectively, and its adoption is expected to bring a clear improvement in responsiveness and overall performance once implemented.
We really appreciate your feedback, as it helps us focus on the areas that matter most to user
Best regards,
----------------------------------------------------------------------------------------------------------
Da scheint nun tatsächlich was zu laufen. Warten wirs ab. 