Re: AW: Re: WG: AW: AW: AW: WG: Re: read performance much slow er than write

"Also sprach Peter T. Breuer:"
> > free on servers during the test without -n on servers with merge
> > requests=512 on client with raid5 24,5MB/s read speed:
> > server1:

BTW ... this is not quite the format of the output of "free" on my machine!


> >              total       used       free     
> > Mem:       1036524     114496     922028          
> > -/+ buffers/cache:      30012    1006512
> > Swap:      1999992          0    1999992

I get:

             total       used       free     shared    buffers     cached
 Mem:        254692     251924       2768         0    13260       38428
 -/+ buffers/cache:     200236      54456
 Swap:       530136      55340     474796

and it is the CACHED information that I really want to see (because we are
discussing reading). The output of meminfo may be more detailed. Heck,
it will be:

MemTotal:       254692 kB
MemFree:          2704 kB
Buffers:         13268 kB
Cached:          38420 kB
SwapCached:       6812 kB
Active:         121860 kB
Inactive:         3028 kB
HighTotal:           0 kB
HighFree:            0 kB
LowTotal:       254692 kB
LowFree:          2704 kB
SwapTotal:      530136 kB
SwapFree:       474796 kB
Dirty:              20 kB
Writeback:           0 kB
Mapped:          98692 kB
Slab:           122496 kB
Committed_AS:   140356 kB
PageTables:       1548 kB
VmallocTotal:   778168 kB
VmallocUsed:      1844 kB
VmallocChunk:   776272 kB

It is differences in that output beween using and not using O_DIRECT
that would be interesting (possibly).


Peter