I want to tell you about a bad experience that made me lose days of work and blocked the mailserver for almost 18 hours
I have my mail server in a private cloud (Aruba Italy).
I can also choose the disks based on performance and having a practical vmware director I assign resources based on needs.
Basically I use two types of storage
1) Mixed SSD and SAS storage with high performance and low latency, redundant via hardware RAID and replicated on twin storage.
2) SAS + NL SAS type storage with medium performance and latency, redundant via hardware RAID and replicated on twin storage.
For the mail server I use the first type, but I had the crazy idea of also using a type 2 disk for logs and message archiving
For 2 years now without problems.
But nothing could be more wrong in hindsight.
If you think about it carefully, at the I/O level, the disk for logs and that for message archiving is the one with the most traffic.
What happened:
The provider decided to upgrade the type 2 infrastructure by adding disks, but this caused the storage to go into degraded mode due to the rebuild, so it slowed down.
The mail server went down, without being able to clear the queues on the disk, Windows no longer responded. I was able to restart it only after I understood the problem and removed the offending disk.
I hope this bad experience of mine is useful to someone