pcp
[Top] [All Lists]

Re: Log compression

To: Nathan Scott <nathans@xxxxxxxxxx>
Subject: Re: Log compression
From: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>
Date: Mon, 27 Dec 2010 10:38:25 +1100
Cc: pcp@xxxxxxxxxxx
In-reply-to: <299040159.360521293058921932.JavaMail.root@xxxxxxxxxxxxxxxxxxxxxx>
References: <299040159.360521293058921932.JavaMail.root@xxxxxxxxxxxxxxxxxxxxxx>
Reply-to: kenj@xxxxxxxxxxxxxxxx
On Thu, 2010-12-23 at 10:02 +1100, Nathan Scott wrote:

> 
> > I've been using the -x option to pmlogger_daily for ages so the
> > regular
> > archives are compressed, but have to be uncompressed manually before
> > use.
> 
> Yep, we use that extensively too ... and the manual uncompress is a
> constant source of complaints from ops folk (and myself) ... it'd be
> awesome if that was automated since its such a common operation.

So, something like this ???

$ pmval -t 1h -i "'1 minute'" -f 4 \
-a /var/log/pcp/pmlogger/bozo/20101219.0.bz2 kernel.all.load

metric:    kernel.all.load
archive:   /var/log/pcp/pmlogger/bozo/20101219.0.bz2
host:      bozo
start:     Sun Dec 19 00:10:11 2010
end:       Sun Dec 19 23:10:11 2010
semantics: instantaneous value
units:     none
samples:   23
interval:  3600.00 sec
00:10:11.564  No values available

                 1 minute 
01:10:11.564       0.0900 
02:10:11.564       0.0900 
03:10:11.564       0.1000 
04:10:11.564       0.1000 
05:10:11.564       0.0000 
06:10:11.564       0.0200 
07:10:11.564       0.0800 
08:10:11.564       0.0000 
09:10:11.564       0.1800 
10:10:11.564       0.1300 
11:10:11.564       0.1100 
12:10:11.564       0.0000 
13:10:11.564       0.0000 
14:10:11.564       0.0300 
15:10:11.564       0.0100 
16:10:11.564       0.0600 
17:10:11.564       0.0000 
18:10:11.564       0.0900 
19:10:11.564       0.1300 
20:10:11.564       0.0500 
21:10:11.564       0.0000 
22:10:11.564       0.0000 

This is a small patch with a tiny footprint on existing code and minimal
risk ... once I make it industrial strength I'll commit it to my tree.

But for xmas grins, the current state of the patch is attached.

Attachment: patch.compress
Description: Text Data

<Prev in Thread] Current Thread [Next in Thread>
  • Re: Log compression, Ken McDonell <=