pcp
[Top] [All Lists]

Re: [pcp] Duplicate Names in the PMNS - seeking consensus

To: Mark Goodwin <mgoodwin@xxxxxxxxxx>, PCP <pcp@xxxxxxxxxxx>
Subject: Re: [pcp] Duplicate Names in the PMNS - seeking consensus
From: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon, 19 Jan 2015 16:44:20 +1100
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <54BC983C.9070905@xxxxxxxxxx>
References: <54BA00E9.7050504@xxxxxxxxxxxxxxxx> <54BC4A6F.1070105@xxxxxxxxxx> <54BC76BD.2090908@xxxxxxxxxxxxxxxx> <54BC983C.9070905@xxxxxxxxxx>
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.3.0
On 19/01/15 16:38, Mark Goodwin wrote:
...
yes it's still a tree and there are no cycles, but I was thinking
more about the subtle change in semantics for pmTraversePMNS(),
where the assumption was each metric in the named subtree would
receive exactly one callback - that would now only be true for
each leaf node / name in the PMNS, not each metric (many names
to one metric). Maybe that's not a problem.

Depending on how you squint, pmTravsePMNS() really only guarantees one callback per leaf node below the root of the search ... this would still be true and different names would (possibly) be returned for metrics with the same PMID.

So, I don't think that is an issue. Most clients would treat this no differently to the same metric being named twice on the command line or in a config file.

we could provide a simple tool to list the duplicate pmIDs in a subtree
for this, and I guess qa tests to verify what the developer was expecting.

Good idea, and it could be used to emit a warning from the Install scripts which would trigger qa failures if this was introduced as a regression in an existing PMDA.

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