pcp
[Top] [All Lists]

Re: [pcp] qa/767 failing most places

To: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Subject: Re: [pcp] qa/767 failing most places
From: Nathan Scott <nathans@xxxxxxxxxx>
Date: Tue, 24 Feb 2015 01:55:26 -0500 (EST)
Cc: PCP <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <411990989.13897673.1424739123526.JavaMail.zimbra@xxxxxxxxxx>
References: <54EB8B14.7070605@xxxxxxxxxxxxxxxx> <1281616127.13879459.1424735391627.JavaMail.zimbra@xxxxxxxxxx> <000d01d04fc9$8375ca60$8a615f20$@internode.on.net> <411990989.13897673.1424739123526.JavaMail.zimbra@xxxxxxxxxx>
Reply-to: Nathan Scott <nathans@xxxxxxxxxx>
Thread-index: AQHch2CjyImo8TZUu0YU/mQFSoNSUwJAQuapnNSBxZAFuvc4bD0Qc+iQ
Thread-topic: qa/767 failing most places
Hi Ken,

----- Original Message -----
> ----- Original Message -----
> > > -----Original Message-----
> > > From: Nathan Scott [mailto:nathans@xxxxxxxxxx]
> > > Sent: Tuesday, 24 February 2015 10:50 AM
> > > ....   Ohhhhh... part of it is
> > > going to be a lack of kernel support for setns(2) - I'll need to get some
> > > way
> > > to detect that built into the QA test.
> > 
> > Check qa 377 or 480 for a (hacky) way to check in a qa script
> > HAVE_SOMETHING
> > that maybe set by configure during the build.
> > 
> 
> Taa.  For this one, I think we can make the existing pmcd.feature.containers
> metric export zero if the needed kernel features were not present/found too,
> and then make the common.check _check_containers() routine use that.

I've pushed this one - I think it will resolve one or two of the cases,
but the "No values" output isn't explained by this ... and I can't seem
to reproduce that on any of my hosts.  Hmm, will ponder some more.

cheers.

--
Nathan

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