> > Peaks for striped are @ 80, peaks for single are ~ 100, peaks for
> > mirror are around 100. I'm curious if hw mirroring would help, as I am
> > about a 4 disk raid 5. But I'm not likely to have the proper drives
> > for that in time to do the testing, and I like the ablity to break the
> > mirror for a backup. For comparison, that same system was doing 20-50
> > without the extra 512 stick of ram and on the internal single drive.
Upon some further poking around, I have determined that there were procedural errors that make the data inconsistent. I
believethat half of the ram cache may not have been in the state that I thought it was in for all of the tests.
> Hm. That's still very low (about the same as a single modern IDE drive).
> I'm looking for an IDE RAID controller that would get me up into the
> 300-500 reads/writes per second range, for 8K blocks. This should not
> be a problem when doing striping across eight disks that are each
> individually capable of about 90 random 8K reads/writes per second.
> (Depending on the size of the data area you're testing on, of course.)
Running some further tests, I'm seeing software striping in the 125 tps peak/100 sustained range, which is about what
I'mgetting from the split WAL/Data mode right now. I'm still seeing about 10% reads, so there's probably some more to
begained with additional system ram.
eric