pcp
[Top] [All Lists]

Re: [pcp] pmdanfsclient errors

To: Martins Innus <minnus@xxxxxxxxxxx>
Subject: Re: [pcp] pmdanfsclient errors
From: Nathan Scott <nathans@xxxxxxxxxx>
Date: Mon, 5 Oct 2015 21:35:27 -0400 (EDT)
Cc: PCP <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <778323742.49282441.1444088240293.JavaMail.zimbra@xxxxxxxxxx>
References: <560D9214.3050508@xxxxxxxxxxx> <778323742.49282441.1444088240293.JavaMail.zimbra@xxxxxxxxxx>
Reply-to: Nathan Scott <nathans@xxxxxxxxxx>
Thread-index: Zlg6E2x4yL6bU5N2Mn0cMJTEVPNNpACAlL9O
Thread-topic: pmdanfsclient errors
Hi Martins,

----- Original Message -----
> ----- Original Message -----
> > Hi,
> >      Commit c683075c3aa39d41d18b2ee35f426c726f7af193 generates errors in
> > pmda startup for my nfs machines.
> > 
> > I get things like the following:
> > 
> > ......
> > Argument "tcp" isn't numeric in numeric ne (!=) at
> > /var/lib/pcp/pmdas/nfsclient/pmdanfsclient.pl line 422.
> > Argument "all" isn't numeric in numeric ne (!=) at
> > /var/lib/pcp/pmdas/nfsclient/pmdanfsclient.pl line 422.
> > Argument "tcp" isn't numeric in numeric ne (!=) at
> > /var/lib/pcp/pmdas/nfsclient/pmdanfsclient.pl line 422.
> > ......
> > 
> > I can take a look next week unless someone gets to it sooner.
> > 
> 
> I think I see the problem, in that the new checks are not catering
> for the case where a value was observed but its a string instead of
> a number - I'll get a patch out shortly, thanks Martins.

Does the attached patch help?  Not 100% sure about the above theory
any more, because the test code does include string metric checking
and that seems to be working OK for me (even without the attached).

cheers.

--
Nathan

Attachment: pmdanfsclient.patch
Description: Text Data

<Prev in Thread] Current Thread [Next in Thread>