Re: [Ganglia-general] rrdcached /gmetad permission problems after update.

2016-02-26 Thread Adrian Sevcenco 
On 02/23/2016 04:57 PM, Vladimir Vuksan wrote:
> Another thing to try is to switch over to using TCP for rrdcached
> connections since that avoids contention on the rrdcached socket and
> should avoid some of the permissions issues. For example I am using
> following options
> 
> OPTS=" -t 60 -w 180 -z 180 -F -s ganglia -m 664 -l 127.0.0.1:9998 -s
> ganglia -m 777 -P FLUSH,STATS,HELP -l unix:/tmp/rrdcached.limited.sock
> -b /var/lib/ganglia/rrds -B -p /var/lib/ganglia/rrdcached.pid
Hi! Thank you for advice! but after switching to tcp now i have a
problem with this :

Feb 26 18:23:55 monitor /usr/sbin/gmetad[16225]: RRD_update
(/export/ganglia/rrds/__SummaryInfo__/multicpu_wio4.rrd): absolute path
names not allowed when talking to a remote daemon

i have this structure for options :
OPTS="-b ${RRDS_DIR} -B -R -F -j ${RRDCACHED_JOURNAL_DIR} -w
${RRDCACHED_WRITE_TIMEOUT} -z ${RRDCACHED_WRITE_DELAY} -f
${RRDCACHED_WRITE_OLDTIMEOUT} \
-t ${RRDCACHED_WRITE_THREADS} \
-s ${RRDCACHED_USER} -m ${SOCKPERMS} -l ${SOCKET1} \
-s ${RRDCACHED_USER_LIM} -m 777 -P FLUSH,STATS,HELP,INFO -l ${SOCKET2}"

RRDCACHED_USER* is ganglia and the PID file is set up in the init file...
in the end i had to return to my old setting and because the graphs are
not seen by ganglia-web if limited sock is used, i had to use the full
sock..

> In the gmetad.conf you can then add
> 
> rrdcached_address 127.0.0.1:9998
this did not worked, i had to use :
RRDCACHED_ADDRESS=127.0.0.1:9998

btw : why is this setting on different file than gmetad.conf?
Couldn't/shouldn't be in gmetad.conf?

Thank you!
Adrian


> 
> Vladimir
> 
> 02/23/2016 u 09:34 AM, Adrian Sevcenco je napisao/la:
>> On 02/22/2016 10:32 PM, Grigory Shamov wrote:
>>> Hi All,
>>>
>>> I have updated Gmond/Gmetad to 3.7.2 on our Ganglia server that uses also
>>> RRDCached. 
>>> It used to work, and configuration didn't  change, but bow metrics do not
>>> get into the graphs anymore.
>>> In the logs there is a lot of messages about permissions, and a new kind
>>> of message about imuxsock thing:
>>>
>>> Feb 22 13:18:40 host /usr/sbin/gmetad[3554]: RRD_update
>>> (/var/lib/ganglia/rrds/Grex/__SummaryInfo__/rx_bytes_ib0.rrd): rrdcached:
>>> Permission denied.
>>> Feb 22 13:18:40 host rsyslogd-2177: imuxsock begins to drop messages from
>>> pid 3554 due to rate-limiting
>>>
>>> Does anyone know how to fix it? Thank you very much in advance!
>> I had the same problem (and partially i still have one) ..
>> So :
>> 1. that "drop messages" message is because of the write errors .. you
>> can ignore that
>>
>> 2. the problem is with rrdcached .. i tried with making part of group
>> ganglia, change ownership of rrds to ganglia:rrdcached but had the same
>> errors ... in the end i had to make the rrds dir 777 and i get rid of errors
>>
>> 3. ganglia web does not function with the rrdcached limited socket so i
>> had to use the same full socket that gmetad use ... i hope that nothing
>> bad will happen but i have no other choice...
>>
>> HTH,
>> Adrian
>>
>>
>>
>> --
>> Site24x7 APM Insight: Get Deep Visibility into Application Performance
>> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
>> Monitor end-to-end web transactions and take corrective actions now
>> Troubleshoot faster and improve end-user experience. Signup Now!
>> http://pubads.g.doubleclick.net/gampad/clk?id=272487151=/4140
>>
>>
>> ___
>> Ganglia-general mailing list
>> Ganglia-general@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/ganglia-general
> 


-- 
--
Adrian Sevcenco, Ph.D.   |
Institute of Space Science - ISS, Romania|
adrian.sevcenco at {cern.ch,spacescience.ro} |
--



smime.p7s
Description: S/MIME Cryptographic Signature
--
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=272487151=/4140___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general

Re: [Ganglia-general] How stop to report a metric for a died java process.

2016-02-26 Thread Vladimir Vuksan 
Hi Jose,

issue is that JMXAgent is not setting the dmax value which by default 
will be set to 0 aka infinity. I haven't looked at JMXAgent in years so 
I don't really remember if that is possible. See if that is possible.

Vladimir


02/26/2016 u 08:06 AM, José Alejandro Camiño Iglesias je napisao/la:
> Hi,
>
> i am working in a project in whose we use Ganglia and JMXAgent to report some 
> metric of the JMX to the
> gmetad to analyse the cluster status.
>
> We noticed that if the java process that is reporting that metrics died, for 
> example by a OutOfMemory due to the charge on that component,
> the metric is still live in the gmetad and with the last value that was 
> reported by the gmond.
>
> I understand that the JMXAgent report the MBean value to the gmond and where 
> the process die gmond is reporting the last value that
> was published by the JMXAgent to the gmetad.
>
> Exist a form to do that the metric isn't reported or the value be a 
> predefined value (void, 0, -1 or similar)?
>
> Thanks,
>


--
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=272487151=/4140
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general

[Ganglia-general] How stop to report a metric for a died java process.

2016-02-26 Thread José Alejandro Camiño Iglesias 
Hi,

i am working in a project in whose we use Ganglia and JMXAgent to report some 
metric of the JMX to the
gmetad to analyse the cluster status.

We noticed that if the java process that is reporting that metrics died, for 
example by a OutOfMemory due to the charge on that component,
the metric is still live in the gmetad and with the last value that was 
reported by the gmond.

I understand that the JMXAgent report the MBean value to the gmond and where 
the process die gmond is reporting the last value that
was published by the JMXAgent to the gmetad.

Exist a form to do that the metric isn't reported or the value be a predefined 
value (void, 0, -1 or similar)?

Thanks,

José Alejandro Camiño Iglesias
———
Escuela Técnica Superior de Ingenieros Informáticos
Universidad Politécnica de Madrid
ja.cam...@alumnos.upm.es



signature.asc
Description: Message signed with OpenPGP using GPGMail
--
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=272487151=/4140___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general