pcp
[Top] [All Lists]

Re: [pcp] Patch: gfs2 PMDA additional metrics for review/inclusion

To: Nathan Scott <nathans@xxxxxxxxxx>
Subject: Re: [pcp] Patch: gfs2 PMDA additional metrics for review/inclusion
From: Paul Evans <pevans@xxxxxxxxxx>
Date: Tue, 04 Jun 2013 14:26:55 +0100
Cc: pcp@xxxxxxxxxxx
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <1691226900.12032333.1369450053711.JavaMail.root@xxxxxxxxxx>
References: <517FBD63.3010804@xxxxxxxxxx> <1117296374.7864618.1367373452615.JavaMail.root@xxxxxxxxxx> <5180D92F.40809@xxxxxxxxxx> <519E43ED.8020701@xxxxxxxxxx> <1691226900.12032333.1369450053711.JavaMail.root@xxxxxxxxxx>
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:17.0) Gecko/20130311 Thunderbird/17.0.4
Hi Nathan,

I've completed another round of changes as based on your comments and fixed the issues with the QA tests:

Removed the needless on-stack "value" local variable in gfs2_store() and all other related code belonging to it.

Fixed and improved both QA tests to exercise the pmda 654 and 655. They have also had their expected outputs recorded into their respective .out files and changes have been made to common.gfs2 to reflect this.

These changes are available at git://github.com/pauljevans/pcp.git gfs2

On 05/25/2013 03:47 AM, Nathan Scott wrote:
The next steps from there would be:

- A man page for pmdagfs2 would be a good idea, esp. since its a fairly
   complex agent - see man/man1/pmdacisco.1 as an example.  Description of
   the store-to-enable/disable mechanism would be helpful.

- A client tool to drive the extraction and reporting of stats from the
   various cluster nodes, manage the enabling / disabling of tracing, etc.
   As discussed, python might be an appropriate language (other option is
   C), some tests and a man page for the new tool would be a good idea as
   well.  Looking forward to this - should be an interesting bit of code!
Looking forward these would be the most logical next steps to make, along with adding additional metrics.

I look forward to feedback.

Regards,

Paul.

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