pcp
[Top] [All Lists]

Re: [pcp] pmcd fails

To: Nathan Scott <nathans@xxxxxxxxxx>
Subject: Re: [pcp] pmcd fails
From: Jeff Hanson <jhanson@xxxxxxx>
Date: Wed, 12 Jan 2011 09:41:49 -0500
Cc: John Hearns <john.hearns@xxxxxxxxxxx>, pcp@xxxxxxxxxxx
In-reply-to: <1045061412.75161294784104502.JavaMail.root@xxxxxxxxxxxxxxxxxxxxxx>
Organization: SGI
References: <1045061412.75161294784104502.JavaMail.root@xxxxxxxxxxxxxxxxxxxxxx>
User-agent: Mozilla/5.0 (X11; U; Linux i686 (x86_64); en-US; rv:1.9.2.13) Gecko/20101207 Lightning/1.0b2 Thunderbird/3.1.7
On 01/11/2011 05:15 PM, Nathan Scott wrote:
> Hi John,
> 
> ----- "John Hearns" <john.hearns@xxxxxxxxxxx> wrote:
> 
>> I have pcp installed on some workstations
>>
>> pmcd fails to start, and the pmcd.log has:
>>
>> ...
> 
> This looks like the problem resolved by the commit below,
> fixed in 3.3.1.  looks like you are running the SGI rpm
> though, I don't know whether there's been an SGI build
> with that fix included yet.

I think it's actually the SLES rpm not an SGI one.  However in
current release of SGI Foundation Software as built for use with Tempo 2.2 we 
have

pcp-3.3.3-1sgi702r2.sles10.x86_64.rpm
pcp-gui-1.4.5-1sgi702r2.sles10.x86_64.rpm
pcp-libs-3.3.3-1sgi702r2.sles10.x86_64.rpm
pcp-libs-devel-3.3.3-1sgi702r2.sles10.x86_64.rpm
pcp-pmda-cluster-1.0.4-1sgi702r2.sles10.x86_64.rpm
pcp-pmda-cluster-client-1.0.4-1sgi702r2.sles10.x86_64.rpm
pcp-pmda-infiniband-1.1.1-1sgi702r2.sles10.x86_64.rpm
pcp-sgi-3.1.1-sgi702r4.sles10.x86_64.rpm

As well as source rpms.  These aren't going to be SLES ones but you could use
the source rpm and rebuild it for your workstations.  Or open a case with Novell
if it's not SGI systems.  If it is your SGI hardware then open a case with us
and reference this email to help the CSC to find what you are looking for.

> 
> commit b8447dbb887a7d14cbcdddd3d8928fcb7838e70e
> Author: Nathan Scott <nathans@xxxxxxxxxx>
> Date:   Tue Jun 29 11:59:13 2010 +1000
> 
>     Fix a sigsegv in pmdalinux in pernode cpu metrics.
>    
>     We need to ensure whenever the cpu indom is updated that
>     the node indom is done too - we were calling the update
>     routines one-after-the-other but missed a spot (instance
>     lookup).  Now its just one call, and it does both things
>     so it cannot ever be missed.
> 
> 
> cheers.
> 
> --
> Nathan
> 
> _______________________________________________
> pcp mailing list
> pcp@xxxxxxxxxxx
> http://oss.sgi.com/mailman/listinfo/pcp
> 


-- 
-----------------------------------------------------------------------
Jeff Hanson - jhanson@xxxxxxx - Field Technical Analyst

You can choose a ready guide in some celestial voice.
If you choose not to decide, you still have made a choice.
You can choose from phantom fears and kindness that can kill;
I will choose a path that's clear
I will choose freewill. - Lee/Lifeson/Peart

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