Well I would gladly supply you my data..
This is what I get when I set $debug=1 in graph.inc.php.
Array ( [0] => bill_bits [type] => bill [1] => bill [subtype] => bits [2] => bits ) SQL[SELECT * FROM `bills` WHERE bill_id = '22'] SQL[] SQL[SELECT delta FROM bill_data WHERE bill_id = '22' AND timestamp> '20130501000000' AND timestamp<= '20130530235959' ORDER BY delta ASC] SQL[SELECT SUM(period) as period, SUM(delta) as total, SUM(in_delta) as inbound, SUM(out_delta) as outbound FROM bill_data WHERE bill_id = '22' AND timestamp> '20130501000000' AND timestamp<= '20130530235959'] SQL[SELECT count(delta) FROM bill_data WHERE bill_id = '22' AND timestamp> '20130501000000' AND timestamp<= '20130530235959'] SQL[SELECT (in_delta / period * 8) AS rate FROM bill_data WHERE bill_id = '22' AND timestamp> '20130501000000' AND timestamp<= '20130530235959' ORDER BY rate ASC] SQL[SELECT count(delta) FROM bill_data WHERE bill_id = '22' AND timestamp> '20130501000000' AND timestamp<= '20130530235959'] SQL[SELECT (out_delta / period * 8) AS rate FROM bill_data WHERE bill_id = '22' AND timestamp> '20130501000000' AND timestamp<= '20130530235959' ORDER BY rate ASC] SQL[SELECT * FROM `bill_ports` AS B, `ports` AS P, `devices` AS D WHERE B.bill_id = '22' AND P.port_id = B.port_id AND D.device_id = P.device_id] OK u:0.16 s:0.00 r:0.17
graph /tmp/yhPqSmYU1s6LIbKi.png --alt-autoscale-max --rigid -E --start 1367359200 --end 1369951199 --width 1050 --height 300 -c BACK#EEEEEE00 -c SHADEA#EEEEEE00 -c SHADEB#EEEEEE00 -c FONT#000000 -c CANVAS#FFFFFF00 -c GRID#a5a5a5 -c MGRID#FF9999 -c FRAME#5e5e5e -c ARROW#5e5e5e -R normal --font LEGEND:8:DejaVuSansMono --font AXIS:7:DejaVuSansMono --font-render-mode normal COMMENT:'Bits/sec Current Average Maximum ' COMMENT:'\l' DEF:in0=/opt/observium/rrd/sit-cre-02-7606s/port-371.rrd:INOCTETS:AVERAGE DEF:out0=/opt/observium/rrd/sit-cre-02-7606s/port-371.rrd:OUTOCTETS:AVERAGE CDEF:inB0=in0,8,* CDEF:outB0=out0,8,* CDEF:outB0_neg=outB0,-1,* CDEF:octets0=inB0,outB0,+ VDEF:totin0=inB0,TOTAL VDEF:totout0=outB0,TOTAL VDEF:tot0=octets0,TOTAL AREA:inB0#B6D14B:'Vlan166 In': GPRINT:inB0:LAST:%6.2lf%sbps GPRINT:inB0:AVERAGE:%6.2lf%sbps GPRINT:inB0:MAX:%6.2lf%sbps 'COMMENT:\n' 'HRULE:0#A0A0E5: Out' GPRINT:outB0:LAST:%6.2lf%sbps GPRINT:outB0:AVERAGE:%6.2lf%sbps GPRINT:outB0:MAX:%6.2lf%sbps 'COMMENT:\n' DEF:in1=/opt/observium/rrd/sit-cre-02-7606s/port-372.rrd:INOCTETS:AVERAGE DEF:out1=/opt/observium/rrd/sit-cre-02-7606s/port-372.rrd:OUTOCTETS:AVERAGE CDEF:inB1=in1,8,* CDEF:outB1=out1,8,* CDEF:outB1_neg=outB1,-1,* CDEF:octets1=inB1,outB1,+ VDEF:totin1=inB1,TOTAL VDEF:totout1=outB1,TOTAL VDEF:tot1=octets1,TOTAL AREA:inB1#A4C445:'Vlan161 In':STACK GPRINT:inB1:LAST:%6.2lf%sbps GPRINT:inB1:AVERAGE:%6.2lf%sbps GPRINT:inB1:MAX:%6.2lf%sbps 'COMMENT:\n' 'HRULE:0#9090D3: Out' GPRINT:outB1:LAST:%6.2lf%sbps GPRINT:outB1:AVERAGE:%6.2lf%sbps GPRINT:outB1:MAX:%6.2lf%sbps 'COMMENT:\n' CDEF:inoctets=in0,UN,0,in0,IF,in1,UN,0,in1,IF,+ CDEF:outoctets=out0,UN,0,out0,IF,out1,UN,0,out1,IF,+ CDEF:doutoctets=outoctets,-1,* CDEF:inbits=inoctets,8,* CDEF:outbits=outoctets,8,* CDEF:doutbits=doutoctets,8,* VDEF:95thin=inbits,95,PERCENT VDEF:95thout=outbits,95,PERCENT VDEF:d95thout=doutbits,5,PERCENT LINE1.25:inbits#005500:'Total In ' GPRINT:inbits:LAST:%6.2lf%s GPRINT:inbits:AVERAGE:%6.2lf%s GPRINT:inbits:MAX:%6.2lf%s GPRINT:95thin:%6.2lf%s\n LINE1.25:doutbits#000055:'Total Out ' GPRINT:outbits:LAST:%6.2lf%s GPRINT:outbits:AVERAGE:%6.2lf%s GPRINT:outbits:MAX:%6.2lf%s GPRINT:95thout:%6.2lf%s\n COMMENT:'\r' HRULE:51244383.26#cc0000:'95th %ile : 51.2Mbps (out) (CDR: 300Mbps)' HRULE:-51244383.26#cc0000 'AREA:outB0_neg#A0A0E5::' 'AREA:outB1_neg#9090D3::STACK' HRULE:0#999999
command returned 0 (1131x467 )
Can you at least give me some pointers, where to look at, what file to compare to a vanilla distribution etc ?
Thanks,
Hynek
On 3.6.2013 18:29, Adam Armstrong wrote:
I can't seem to replicate what you're seeing. My test data is fine.
adam.
On 2013-06-03 17:23, Hynek Med wrote:
but why / how to fix that?
h.
On 3.6.2013 18:00, Adam Armstrong wrote:
Sounds like it's calculating averages.
adam.
On 2013-06-03 16:10, Hynek Med wrote: Probably somewhere between r3465 to r3981 (yes, I know) billing started producing wrong results. now at 4102, the situation remains the same. I do run the discovery script after each update.
Graphs (quick graphs) look fine with real traffic accounted, but the calculated P95 is unrealistically low (something like 1/3 of the real P95, but this ratio is different for each case). Both with single VLAN and multiple VLAN bills. One of the bills (with id=1) works completely fine though.
Any ideas where to look at / what to compare at the working case with the non-working case?
Thanks,
Hynek _______________________________________________ observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium
observium mailing list observium@observium.org http://postman.memetic.org/cgi-bin/mailman/listinfo/observium