[ic] CPU upgrade for IC server?
steven at endpoint.com
Wed Jun 17 13:38:11 UTC 2009
>>> Does this mean I'm alright on memory?
>>> Mem: 1028780k total, 709692k used, 319088k free, 55036k buffers
>> Looks reasonable to me, assuming that's what it typically looks like and
>> you're not actively swapping as shown by "si"/"so" columns in vmstat.
> I think this is more typical:
> Mem: 1028780k total, 974664k used, 54116k free, 226184k buffers
> Swap: 2008116k total, 31916k used, 1976200k free, 326828k cached
> procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
> r b swpd free buff cache si so bi bo in cs us sy id wa
> 1 0 31916 54164 226184 326832 1 1 14 3 23 26 35 11 54 0
> Do the "1" in si and so mean I'm actively swapping?
Technically, yes; however, swapping in 1 page and swapping out 1 page is not
enough of a pattern to be concerned about. Notice your cpu columns: the 'wa'
column is zero, so you're not waiting for IO to occur, yet the 'us' column is
35, so you're spending quite a bit of time in your application (i.e., the
'userspace' part of code), yet the overall system is 54% idle.
A suggestion: install and run sar (the 'sysstat' package on a RH system; I
believe it is 'atsar' on Debian). That will do data gathering on a regular
basis so that you can then analyze (and correlate) with problems.
End Point Corporation
More information about the interchange-users