[rrd-users] Not a simple unsigned integer

[Paul Watson]

ERROR: stats_01255.rrd: not a simple unsigned integer: '-3220958'
rrdtool update stats_01255.rrd
1334893211:6211062:1222:0:948:-3220958:2189:703:507::0

ERROR: stats_01342.rrd: not a simple unsigned integer: '-6751407'
rrdtool pinn_db/stats_01342.rrd
1334893211:244060:351:72:0:-6751407:436:119:5::73

ERROR: stats_01511.rrd: not a simple unsigned integer: '-2870309'
rrdtool stats_01511.rrd 1334893211:4435410:257:0:0:-2870309:266:216:7::0

However prior to these errors

11:00:29   164  Processing Stats Timestamp 1334892616:164
11:00:33   676  Processing Stats Timestamp 1334892619:676
11:00:52   779  Processing Stats Timestamp 1334892901:779
11:01:13    20  Processing Stats Timestamp 1334892905:20
11:01:13   162  Processing Stats Timestamp 1334892908:162
11:01:17  2902  Processing Stats Timestamp 1334892911:2902
11:02:27   164  Processing Stats Timestamp 1334892915:164
11:02:33   676  Processing Stats Timestamp 1334892918:676
11:02:51   779  Processing Stats Timestamp 1334893202:779
11:03:09    20  Processing Stats Timestamp 1334893205:20
11:03:10   162  Processing Stats Timestamp 1334893208:162
11:03:14  2902  Processing Stats Timestamp 1334893211:2902

The 'erroring' RRDs will have been processed successfully in those batches -
at least twice (they are all in the 2902 batches)

The data cannot (ok should not) change from positive to negative in the 5
min time window we are looking at.

Guess I'm heading to the data sources now and find how they are becoming
negative

Thanks for all your help

Cheers
Paul

-----Original Message-----
From: Tobias Oetiker [mailto:tobi at oetiker.ch] 
Sent: Friday, April 20, 2012 10:46 AM
To: Paul Watson
Cc: rrd-users at lists.oetiker.ch
Subject: RE: [rrd-users] Not a simple unsigned integer

Hi Paul,

Today Paul Watson wrote:

> The same update is called everytime,
>
> rrdtool update ${STUBDIR}/stats_${ID}.rrd
>
${EPOCH}:${SELECT}:${INSERT}:${UPDATE}:${DELETE}:${BREAD}:${BWRITE}:${DREAD}
> :${DWRITE}:${RATIO}:${SEQSCANS}
>
> All the data values come from the database where the columns are defined
as
> int8 (8 byte ints).  Their values can go negative if we wrap past the int8
> max, if the database is restarted then the values start again from zero
>
> The rrd is defined as
>
> rrdtool create ${FNAME}            \
>             --start ${EPOCH}               \
>             DS:select:COUNTER:600:U:U      \
>             DS:insert:COUNTER:600:U:U      \
>             DS:update:COUNTER:600:U:U      \
>             DS:delete:COUNTER:600:U:U      \
>             DS:bread:COUNTER:600:U:U       \
>             DS:bwrite:COUNTER:600:U:U      \
>             DS:dread:COUNTER:600:U:U       \
>             DS:dwrite:COUNTER:600:U:U      \
>             DS:ratio:COUNTER:600:U:U       \
>             DS:seqscans:COUNTER:600:U:U    \
>             RRA:AVERAGE:0.5:1:600          \
>             RRA:AVERAGE:0.5:6:700          \
>             RRA:AVERAGE:0.5:24:775         \
>             RRA:AVERAGE:0.5:288:797
>
> They are 4 distinct rrd templates but they all follow the above 'style'. I
> have around 70K rrds defined at the moment, of which maybe 2-300 show
> 'problems'

so add some print statements to your code to create an output of the

* the rrdtool create sequence to cretate one of the problem rrds
* the rrdtool update sequence necessary to reproduce the problem

pack it all in a shell script and let me have it ...

thanks
tobi



>
> Cheers
> Paul
>
>
> -----Original Message-----
> From: Tobias Oetiker [mailto:tobi at oetiker.ch]
> Sent: Friday, April 20, 2012 10:27 AM
> To: Paul Watson
> Cc: rrd-users at lists.oetiker.ch
> Subject: RE: [rrd-users] Not a simple unsigned integer
>
> Hi Paul,
>
> Today Paul Watson wrote:
>
> > True but .....
> >
> > Having reset ALL the rrds and then ran the raw data through successfully
> > with rrdtool 1.3.8 I then reset again and ran with rrdtool 1.4.7 and got
> the
> > following
> >
> > stats_00093.rrd: not a simple unsigned integer: '-4997672'
> > stats_00095.rrd: not a simple unsigned integer: '-6556514'
> > stats_00109.rrd: not a simple unsigned integer: '-5429246'
> > stats_00166.rrd: not a simple unsigned integer: '-1904720'
> > stats_00167.rrd: not a simple unsigned integer: '-6629670'
> > stats_00172.rrd: not a simple unsigned integer: '-3321973'
> > stats_00178.rrd: not a simple unsigned integer: '-5320095'
> > stats_00179.rrd: not a simple unsigned integer: '-7117550'
> > stats_00253.rrd: not a simple unsigned integer: '-3944869'
> > stats_00300.rrd: not a simple unsigned integer: '-6854738'
> >
> > Note the rrds are number sequentially and the problem only affects  some
> > rrds, all are defined that same. I am NOT passing in negative data, and
> ALL
> > the data is taken from a database where the columns are defined as
integer
> > (8 bytes).
>
> then which rrdtool update calls are causing this error to appear ?
>
> cheers
> tobi
>
>

-- 
Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten, Switzerland
http://it.oetiker.ch tobi at oetiker.ch ++41 62 775 9902 / sb: -9900