When you insert data into a database and run COMMIT you expect things to be there: Atomically, Consistent, Isolated and Durable, like Codd commanded us 40 years ago, but also quickly. There is a surprising amount of sophistication being poured into this, but since I do not want to shame MongoDB and Redis developers in this post, I am not going to talk about that much in this place.

We are instead trying to understand what our databases are doing all day, from the point of view of the storage stack.

High-end hardware at sensational prices! Our cloud servers are located in our own state-of-the-art data centers and provide top performance with their Intel Skylake Xeon CPUs and speedy NVMe SSDs for all your resource-intensive projects.

To give you a little bit of background when the OSD writes into his journal it uses D_SYNC and O_DIRECT. Writing with O_DIRECT bypasses the Kernel page cache, while D_SYNC ensures that the command won’t return until every single write is complete. So yes, basically the OSD forces all the writes to be flushed prior to start the next IO.

EnhanceIO can work with any block device, be it an entire physical disk, an individual disk partition, a RAIDed DAS device, a SAN volume, a device mapper volume or a software RAID (md) device.