pcp
[Top] [All Lists]

Re: Multi-Volume Archive + Live Data Playback for PCP Client Tools

To: Dave Brolley <brolley@xxxxxxxxxx>
Subject: Re: Multi-Volume Archive + Live Data Playback for PCP Client Tools
From: fche@xxxxxxxxxx (Frank Ch. Eigler)
Date: Sun, 02 Nov 2014 10:48:34 -0500
Cc: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>, "'PCP Mailing List'" <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <545110DC.2020104@xxxxxxxxxx> (Dave Brolley's message of "Wed, 29 Oct 2014 12:07:56 -0400")
References: <542C21AE.1010504@xxxxxxxxxx> <007e01cfe010$7867f090$6937d1b0$@internode.on.net> <545110DC.2020104@xxxxxxxxxx>
User-agent: Gnus/5.1008 (Gnus v5.10.8) Emacs/21.4 (gnu/linux)
kenj wrote:

> [...]  One simple solution that might be acceptable for 95% of the
> cases would be to rule all of the metadata data differences (except
> instance domains) to be unsupported.  [...]

It'd be interesting to assess how many actual libraries of archives
would meet or fail this test.  Identical "pmdumplog -d"?

A set of outputs from an constantly-configured configured pmlogger may
pass OR fail, depending on whether the pmlogger ran long enough to
sample the entire set of metrics at least once.  (An early SIGTERM
might let only the "log once" metrics show up in the .meta file.)  Or
the order of first-appearance between metrics might be different.  Or
pmcd might have had a version update, and hinv.*.online metrics may
have appeared from one day to the next.

- FChE

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