In that case you might want to look if your SSDs support lower power states
is there empirical evidence / actual measurements that the use of ssd makes a meaningful difference?
i tend to think that no ssd is going to match ram which zfs uses as a buffer.
thoughts?
is there empirical evidence / actual measurements that the use of ssd makes a meaningful difference?
answering my own question: thinking about it, I have a (highly unscientific) observation to the opposite. I have a NAS (not true nas) that runs off of a usb stick (~7MB/s transfer rate).
In real life uses, it is indistinguishable to the same system if I were to run off a SSD or a hard drive, aside from slightly longer boot-up time.
Do you mean (as your second post suggests) if SSDs make a performance difference as boot drives for TrueNAS?
Once the system is up I don’t think there’d be a noticable difference while using the system compared to an HDD.
It does make a difference reliability-wise compared to a USB stick off course.
I’d also think that some UI operations do profit a lot from a fast boot drive - the audit database is stored on the boot disk for exmaple.
Now that I think about it I don’t really know, why it’s not on the system dataset.