Virtual processor usage wrong?

Hi,

(No AIX section?)

I have been collecting load data from uptime on our AIX and Linux servers following Earl Jew's CSoW measurements in his Starting Tactic presentation.

Sadly, lpar2rrd shows at least double the amount of virtual processors as my/Earl's calculations, which is frustrating as I have a hard time getting my capacity planning paradigm through.

I had a look at lpar2rrd.pl and found this:

sub get_smt_details {

 my ( $wrkdir, $server, $lpar ) = @_;

 my $server_space = $server;

 my $lpar_space  = $lpar;

 # check if exists data/server/hmc/lpar/cpu.txt where the OS agent saves SMT

But all the cpu.txt file I have found are wrong:

# cat cpu.txt

4

On the actual LPAR:

# smtctl

This system is SMT capable.

This system supports up to 8 SMT threads per processor.

SMT is currently enabled.

SMT boot mode is set to enabled.

SMT threads are bound to the same virtual processor.

proc0 has 8 SMT threads.

Bind processor 0 is bound with proc0

Bind processor 1 is bound with proc0

Bind processor 2 is bound with proc0

Bind processor 3 is bound with proc0

Bind processor 4 is bound with proc0

Bind processor 5 is bound with proc0

Bind processor 6 is bound with proc0

Bind processor 7 is bound with proc0

I think this accounts for the virtual processor graph being doubled. I'm not the greatest Perl reader though, so struggling a little bit debugging this.

I can't figure out where lpar2rrd actually try to get the SMT setting from or maybe it defaults to 4 because it can't, but I can't find that either.

Regards,

Henrik Morsing