pcp
[Top] [All Lists]

Re: [pcp] ganglia2pcp question

To: Martins Innus <minnus@xxxxxxxxxxx>
Subject: Re: [pcp] ganglia2pcp question
From: Nathan Scott <nathans@xxxxxxxxxx>
Date: Thu, 5 Feb 2015 18:21:43 -0500 (EST)
Cc: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>, pcp@xxxxxxxxxxx
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <54D3CD6D.3060303@xxxxxxxxxxx>
References: <54D0FC34.6070401@xxxxxxxxxxx> <54D12955.1050805@xxxxxxxxxxxxxxxx> <54D3CD6D.3060303@xxxxxxxxxxx>
Reply-to: Nathan Scott <nathans@xxxxxxxxxx>
Thread-index: y7iGSkfD0N1UIrhcQNcK2SQbHau2nA==
Thread-topic: ganglia2pcp question
Hi Martins,

----- Original Message -----
> "The hostname and timezone will default to the local host, but can be
> explicitly set after pmiStart() by calling pmiSetHostname() or
> pmiSetTimezone(). "
> 
> But the examples I looked at (mover2pcp, iostat2pcp) explicitly set the
> timezone to UTC and only change it based on the presence of a command
> line option, ignoring, I think, what may be configured on the local
> host.
> 
> Is that the way to go for default options?  In our use case we'd
> always be setting an option for local time.
> 
> The ganglia stats are already in unix time, so no conversion is needed
> there, but this would affect parsing start end/times, etc I think.

If you know for certain what the timezone on the host from which
the metrics were collected is, its best to default to storing
that in the log label.  In the iostat case at least, this is not
known though, IIRC.  That data is often imported from text output
sent from a customer system, without timezone info embedded in it.

cheers.

--
Nathan

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