3 ms·
If I recall correctly, Oracle best practices back in the day (waaaay back) was always put the DB on raw partitions, which caused us some issues with existing to
by kjs3 25d ago
If I recall correctly, Oracle best practices back in the day (waaaay back) was always put the DB on raw partitions, which caused us some issues with existing tooling. I've been fortunate enough to not have to pay attention to Oracle best practices in quite a while.
- angry_octet 25d agoHigh performance databases still use raw I/O. You get much better control of data placement and write ordering, without filesystem confounders. With traditional spinning rust being replaced with SSDs and SMR it can be even more important. You may even have different firmware on the disks / disk controllers for database disks.
- kjs3 23d agoSure...pretty clear there's a use case. I just recall our ops folks having to retool some of their playbooks because of the assumption there would be a standard filesystem on every active slice. Fortunately, the whole "load a particular firmware on the disks" wasn't new. That was something some of the HW RAID vendors recommended.