[rrd-users] graph from mysql

Thu May 16 19:14:26 CEST 2013

aah really it returns 64 bit counters, i had no idea about this, can u tell
me the logic i mean wat other oids those are 32bit i need to add HC to it ?

thanks,

On Thu, May 16, 2013 at 8:40 PM, Martin Sperl <rrdtool at martin.sperl.org>wrote:

> /derive does essentially take the "difference" of 2 consecutive "bins" -
> say 5 minute intervals - and if the value is way off (>2^31) , then it
> assumes it is a wraparound and "corrects" for that.
>
> as for SNMP V2: you need to use a different OID to get the corresponding
> 64 bit values - really: look on the internet to get those OID numbers in
> question.
> (a quick look shows: snmpget -v 2c -c public 14.32.5.18 ifHCInOctets.1
> could do the trick for you - maybe you need to modify it for your needs)
>
> Martin
>
> On 15.05.2013, at 17:34, Shabbir Ahmed wrote:
>
> thanks for your email, before ur msg i tried my self /derive and the
> graphs are normal. kindly look at attachment.
>
> actually i fetch data from switches and i have tried option -v 2c but the
> result is same, 32 bit counter. can u explain it to me wat derive does and
> i dont have to worry about other thigns like calculating data etc my self,
> if you dont mind.
>
> can u also look at my graph generating cmd to verify if im not doing any
> thing wrong,
>
> rrdtool graph test.png --imgformat=PNG --start=-2day --end=-1day
> -S=300    "DEF:up=$UP:min:AVERAGE"     "DEF:down=$DOWN:min:AVERAGE"
> AREA:down#2ECCFA:"Out traffic"   AREA:up#013ADF:"Download"
> --vertical-label=bytes/sec --title="ADSL bandwidth graphs"
> GPRINT:up:LAST:"Current\:%8.2lf %s\n"  GPRINT:up:AVERAGE:"Average\:%8.2lf
> %s\n"  GPRINT:up:MAX:"Maximum\:%8.2lf %s\n" --font TITLE:12: --font AXIS:8:
> --font LEGEND:10: --font UNIT:8: AREA:up#013ADF:"In traffic"
>
>
> I am sorry if im teasing you,
> Shabbir.
>
>
> On Wed, May 15, 2013 at 7:36 PM, Martin Sperl <rrdtool at martin.sperl.org>wrote:
>
>> so you are using the 32bit counter.
>>
>> Then if you insert into the database, you:
>> either have to do the "wraparound" logic yourself prior to inserting the
>> values
>> or you insert the "absolute" counter values and use /derive (as explained
>> before)
>> if you use the rrd flatfile database via rrdcreate and update, then
>> (depending on the definition of the rrd-file) rrd will try to do the
>> "overflow" detection...
>>
>> Martin
>>
>> P.s: the 64bit counter are typically only available if you use snmp V2C -
>> google for "64 bit counter SNMP" and you should find enough information on
>> how to use it.
>>
>>
>> On 15.05.2013, at 10:07, Shabbir Ahmed wrote:
>>
>> quering the device snmpbulkget -c public 119.156.203.217 .
>> 1.3.6.1.2.1.2.2.1.10 -v 2c
>> returns 32bit integers
>> IF-MIB::ifInOctets.20971520 = Counter32: 0
>> IF-MIB::ifInOctets.20972032 = Counter32: 0
>> IF-MIB::ifInOctets.20972544 = Counter32: 1295822366
>> IF-MIB::ifInOctets.20973056 = Counter32: 1295822366
>> IF-MIB::ifInOctets.20973568 = Counter32: 1295822366
>> IF-MIB::ifInOctets.20974080 = Counter32: 1295822366
>>
>>
>> 2nd question is most of the values are always same is it the reason port
>> is down?
>>
>> 3rd question what if counter reaches its up limit and starts from 0
>> again, how rrd tool manages it? or should we have to manage our self?
>>
>> kindly help further with this issue.
>>
>> thanks,
>>
>>
>> On Wed, May 15, 2013 at 11:53 AM, Martin Sperl <rrdtool at martin.sperl.org>wrote:
>>
>>> Seems as if you are having "integer overflow" issues reading the 32bit
>>> SNMP values (or when calculating the delta of the counters in your program).
>>>
>>> These rollovers is something your application will need to figure out
>>> prior to inserting into the database - either by detecting a "rollover"
>>> (the "rrdtool update" command for rrd files - not used with the dbi backend
>>> - has a provision for that) or by querying the 64bit values on the router
>>> instead and then inserting those values. Alternatively the dbi backend can
>>> also "try" to do the calculations for you on the fly if you insert the
>>> absolute counter values - see the man-page for details of which option to
>>> use... (/derive after the data column)
>>>
>>> Martin
>>>
>>>
>>>
>>> On 14.05.2013, at 20:03, Shabbir Ahmed wrote:
>>>
>>> yes i used, this time i made graphs for my dsl router at home and they
>>> are doing better, but the connection is 4mb and the graph shows 1.5G, i hv
>>> checked database has large values, but these are ported from device, what i
>>> do i need to do now devid the values on something to get the desired value?
>>>
>>>
>>>
>>>
>>> On Tue, May 14, 2013 at 12:31 AM, Martin Sperl <rrdtool at martin.sperl.org
>>> > wrote:
>>>
>>>> Did you try the /rrdfillmissing=1/ as mentioned below as a workaround
>>>> for your "missing" data?
>>>>
>>>> Martin
>>>>
>>>>
>>>>
>>>> On 13.05.2013, at 18:27, Shabbir Ahmed wrote:
>>>>
>>>> some time when i query for snmp the get times out so i cant fill in the
>>>> table at that time, to fix this wat should i do ?
>>>>
>>>> 12:00 i queried db response recorded into mysql
>>>> 12:05 i queried but snmp get timeout, no entry into mysql
>>>> 12:06 i queried and got the response into mysql.
>>>>
>>>>
>>>>
>>>>
>>>> On Mon, May 13, 2013 at 11:36 AM, Martin Sperl <
>>>> rrdtool at martin.sperl.org> wrote:
>>>>
>>>>> Hi!
>>>>>
>>>>> One observation: actually you should see 4 lines not 2 or 1.
>>>>>
>>>>> The other is that your data is possibly not stored "equidistant", so
>>>>> that there are slight "gaps" of say 302 seconds between consecutive
>>>>> datapoints, which would mean that you are producing times like
>>>>> this: 0,302,603,...,3298,3601,... (so there would be a gap at time 3300 to
>>>>> 3599)
>>>>>
>>>>> For debugging purposes: can you please run rrdtool like this:
>>>>> RRDDEBUGSQL=1 *rrdtool graph test.png --imgformat=PNG --start=-1day
>>>>> --width=600 --height=80    -S=300     "DEF:min=$UP:min:AVERAGE"
>>>>> "LINE1:min#336600:value"*
>>>>>
>>>>> Then you will see the SQL that gets executed and then you can run the
>>>>> SQL on your own to understand how this "gaps" are generated from the
>>>>> database...
>>>>>
>>>>> But there is also an option to "fill in" "missing" gaps in rrd via the
>>>>> "/rrdfillmissing=" option.  The idea here is similar to the "xff" argument
>>>>> to rrdcreate for the RRA definition.
>>>>>
>>>>> So you may try to add it to your DEF like this: UP="sql//mysql/host=
>>>>> 127.0.0.1/dbname=newdsl/username=root/password=nocb4enter<http://127.0.0.1/dbname=newdsl/username=root/password=nocb4enter//g3412306061/time/up/idx=201409600>
>>>>> /rrdfillmissing=1//g3412306061/time/up/idx=201409600<http://127.0.0.1/dbname=newdsl/username=root/password=nocb4enter//g3412306061/time/up/idx=201409600>
>>>>> "
>>>>>
>>>>> Ciao,
>>>>> Martin
>>>>>
>>>>> On 12.05.2013, at 15:05, Shabbir Ahmed wrote:
>>>>>
>>>>> Hi, i collect data from switches and save them into mysql,
>>>>> (snmpidx,time,up,down)
>>>>>
>>>>> now i want to draw graphs using
>>>>>
>>>>> DOWN="sql//mysql/host=
>>>>> 127.0.0.1/dbname=newdsl/username=root/password=nocb4enter//g3412306061/time/down/idx=201343232
>>>>> "
>>>>> UP="sql//mysql/host=
>>>>> 127.0.0.1/dbname=newdsl/username=root/password=nocb4enter//g3412306061/time/up/idx=201409600
>>>>> "
>>>>>
>>>>> *rrdtool graph test.png --imgformat=PNG --start=-1day --width=600
>>>>> --height=80    -S=300     "DEF:min=$UP:min:AVERAGE"
>>>>> "LINE1:min#336600:value"     "DEF:avg=$UP:avg:AVERAGE"
>>>>> "LINE1:avg#00FF11:average"     "DEF:max=$UP:max:AVERAGE"
>>>>> "LINE1:max#FF0000:max" "DEF:dw=$DOWN:min:AVERAGE" "LINE1:dw#0033CC:value"
>>>>>
>>>>> *
>>>>> *1. the graphs drawn has gaps (i collect data every five mins) and
>>>>> draw them with 5 mins step but still the gape bw two occurences are empty.
>>>>>
>>>>> *
>>>>> *2. it doesnt draw two lines.
>>>>>
>>>>> *
>>>>> *Kindly help,
>>>>> *
>>>>> *
>>>>> *
>>>>> _______________________________________________
>>>>> rrd-users mailing list
>>>>> rrd-users at lists.oetiker.ch
>>>>> https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>> <test.png>
>>>
>>>
>>>
>>
>>
> <test.png>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.oetiker.ch/pipermail/rrd-users/attachments/20130516/a150304d/attachment-0001.htm 