pcp
[Top] [All Lists]

Re: pcp2rrdtool qa/926 tweak

To: Nathan Scott <nathans@xxxxxxxxxx>
Subject: Re: pcp2rrdtool qa/926 tweak
From: Martins Innus <minnus@xxxxxxxxxxx>
Date: Fri, 06 Mar 2015 10:18:08 -0500
Cc: PCP <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <1426675409.22052119.1425615446159.JavaMail.zimbra@xxxxxxxxxx>
References: <1426675409.22052119.1425615446159.JavaMail.zimbra@xxxxxxxxxx>
User-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.10; rv:31.0) Gecko/20100101 Thunderbird/31.5.0
Nathan,

On 3/5/15 11:17 PM, Nathan Scott wrote:
Hi Martins,

I'm getting reports of failures of qa/926 on exotic hardware -
e.g. s390x and ppc64 - with output like:

ERROR in RRDs::info(./ganglia/gangliatest.64/load_fifteen.rrd) : This RRD was 
created on another architecture
Hmm, OK. I'm note sure what the ability to have common files is. I did a quick search and couldn't find any good docs.

Also, is there a recipe for gathering this ganglia test data?  I
could extract data from some of these hosts, and extend the test
coverage here.  We could perhaps put this recipe in the comment at
the head of the test too, for anyone else wanting to do this later.

The original data just came from a long gone VM where I installed ganglia and let it run for a minute or so and collect data in it's default config. But you can dump this to XML and then re-import on a new arch, to create a set of files with the same metrics. I used that process to create the 32 bit data. Basically:

<on a working host>
rrdtool dump .....

<copy the files to the non working host>
rrdtool restore ......

Fix coming shortly.

Martins

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