From: Green, Simon (Simon.Green@EU.ALTRIA.COM)
Date: Fri May 23 2003 - 10:35:20 EDT
Surely that was only the case on pre 4.2 systems: a fault in the algorithm
for SMP systems. These days, if it reports wait time, then something really
is waiting.
Moving from a 4-way to a 2-way system might lead to increased wait times.
Even though the total CPU capacity is increased there are now only two
engines available, so if one of them gets clogged up servicing I/O you lose
50% of capacity instead of just 25%.
Simon Green
Altria ITSC Europe s.a.r.l.
AIX-L Archive at http://marc.theaimsgroup.com/?l=aix-l&r=1&w=2
AIX FAQ at http://www.faqs.org/faqs/aix-faq/
N.B. Unsolicited email from vendors will not be appreciated.
> -----Original Message-----
> From: Jim Lane [mailto:jimlane@SYMPATICO.CA]
> Sent: 23 May 2003 15:01
> To: aix-l@Princeton.EDU
> Subject: Re: performance question
>
>
> IO wait in and of itself may not be much of a problem. It is
> only counted if
> the CPU wasn't otherwise busy. You seem from your vmstat
> output to have a lot
> of idle time, I'd be a lot more worried if that was near zero
> with high IO
> wait. IO wait can seem high if you have activity spread
> thinly over a lot of
> physical disks. If any one of them is busy when the CPU is
> idle then it
> counts as IO wait.
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 22:16:51 EDT