pcp
[Top] [All Lists]

Re: [pcp] signal pmlogger

To: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>, "'Frank Ch. Eigler'" <fche@xxxxxxxxxx>
Subject: Re: [pcp] signal pmlogger
From: Martins Innus <minnus@xxxxxxxxxxx>
Date: Mon, 14 Apr 2014 15:27:17 -0400
Cc: 'PCP Mailing List' <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <01c401cf55c9$d4322fc0$7c968f40$@internode.on.net>
References: <53482BC5.6070203@xxxxxxxxxxx> <y0mr453663h.fsf@xxxxxxxx> <01c401cf55c9$d4322fc0$7c968f40$@internode.on.net>
User-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7; rv:24.0) Gecko/20100101 Thunderbird/24.4.0
Ken/Frank,
Thanks, I played with pmlc a little bit and have a couple more questions. I may have simplified my initial description too much. Assume a setup where we have a manually crafted pmlogger config with hundreds of metrics being logged over several different time intervals: Some at 1 minute, 10 minutes and 30 minutes.

If I start some compute process, I want to make sure that I log all of the above described metrics at least once over the life of that process, but otherwise keep the logging schedule unchanged. If the process runs for less than 30 minutes, currently I will get anywhere from a subset to none of the metrics logged, depending on the length of the compute process.

    Playing with pmlc, i can do a:

log mandatory on once <list of all my existing metrics>

to force a logging event when the compute process starts, but then the logger stops its regular logging schedule. So I assume then I have to write a script that parses the logger config file and re-enables the logging as it was with pmlc commands? I didn't see a save/load config set of commands from pmlc unless I missed it.

I can go down that path if that is the right way to do it. But I was hoping for some sort of "log this list of metrics right now, just once, but don't otherwise change pmlogger" command.

This is all on the local host running just a primary logger.

pmie support down the road might be interesting, but for now I know when I want these logging events to occur.

Thanks

Martins

On 4/11/14 5:05 PM, Ken McDonell wrote:
Frank's correct.  pmlc was conceived at the same time pmlogger was created
for just this purpose ... and the equally important case - after some
elapsed time, if we don't know what's going on, we're never going to find
out, so stop logging those additional metrics.

All of the plumbing already exists for both the turning on and turning off
to be driven by pmie (if the triggers to stop/start can be captured by pmie
rules).  It is all a bit low-level, but has been made to work in the past
and will still work.

If you need any assistance putting this together, I can help if I know the
specific details.

-----Original Message-----
From: pcp-bounces@xxxxxxxxxxx [mailto:pcp-bounces@xxxxxxxxxxx] On
Behalf Of Frank Ch. Eigler
Sent: Saturday, 12 April 2014 5:04 AM
To: Martins Innus
Cc: PCP Mailing List
Subject: Re: [pcp] signal pmlogger


minnus wrote:

[...]  I'm trying to find a way to send a signal to pmlogger to log
certain metrics immediately. [...]
See the pmlc tool.  It can direct a pmlogger instance to log new metrics
on
demand.  (One can imagine a future where this is automated via pmie or
other ways.)

- FChE

_______________________________________________

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