Re: [pcp] Invalid proc.memory.vmhwn metric

[Nathan Scott <nathans@xxxxxxxxxx>]

Hi Marko,

----- Original Message -----
> Hi,
> 
> From http://man7.org/linux/man-pages/man5/proc.5.html:
> 
>   * VmHWM: Peak resident set size ("high water mark").
> 
> There's no proc.memory.vmhwm available as one would expect but
> proc.memory.vmhwn - and it turns out that the same typo is present
> everywhere in PCP sources so the value will always be zero.

Good catch.

> How should this be corrected? I could have sent a simple patch changing
> hwn->hwm everywhere

Yep, that's what's needed here.  (could you send that patch through?)

> but are there any backward compatibility
> restrictions that should be taken into account?

Noone can be validly using the "vmhwn" metric, so I vote to switch it
(correct error handling will kick in next release via 'unknown metric
name' return code).

We could write a pmlogrewrite(1) rule, but that seems like overkill in
this situation.

cheers.

--
Nathan