MRTG log stores different data

MRTG log stores different data

Post by Vodochn » Thu, 17 Jul 2003 02:47:29



Can anyone help me to understand why MRTG log store different data?

Monitored parameter "duration" inserted into the database table every
5 min. MRTG executed one minute later. It generates ARCOpenCase.log
and graphs. It works fine but log does not show real data.
Look at the top 20 records from the source table and log. I would
expect they all shoud match but only some of them are.

Comparing values in source table and last column in MRTG log :
1058286061 266          1058286000 264 264 266 266
1058285761 263          1058285700 263 263 263 263
1058285461 263          1058285400 394 394 576 576      <- ?
1058285161 576          1058285100 708 708 890 890      <- ?
1058284861 890          1058284800 600 600 890 890      
1058284562 200          1058284500 345 345 546 546      <- ?
1058284261 546          1058284200 403 403 546 546
1058283961 206          1058283900 277 277 376 376      <- ?
1058283661 376          1058283600 303 303 376 376
1058283361 203          1058283300 223 223 250 250      <- ?
1058283061 250          1058283000 242 242 250 250
1058282761 233          1058282700 227 227 233 233
1058282461 220          1058282400 212 212 220 220
1058282162 203          1058282100 210 210 220 220      <- ?
1058281861 220          1058281800 212 212 220 220

Source table:
date                      timestamp   duration
------------------------- ----------- ---------
2003-07-15 09:26:01.557   1058286361  280
2003-07-15 09:21:01.400   1058286061  266        
2003-07-15 09:16:01.417   1058285761  263        
2003-07-15 09:11:01.400   1058285461  263        
2003-07-15 09:06:01.510   1058285161  576        
2003-07-15 09:01:01.823   1058284861  890        
2003-07-15 08:56:02.073   1058284562  200        
2003-07-15 08:51:01.463   1058284261  546        
2003-07-15 08:46:01.417   1058283961  206        
2003-07-15 08:41:01.433   1058283661  376        
2003-07-15 08:36:01.463   1058283361  203        
2003-07-15 08:31:01.463   1058283061  250        
2003-07-15 08:26:01.683   1058282761  233        
2003-07-15 08:21:01.637   1058282461  220        
2003-07-15 08:16:02.213   1058282162  203        
2003-07-15 08:11:01.433   1058281861  220        
2003-07-15 08:06:01.417   1058281561  203
2003-07-15 08:01:01.387   1058281261  216
2003-07-15 07:56:01.387   1058280961  216
2003-07-15 07:51:01.527   1058280661  186

ARCOpenCase.log :
1058286426 280 280
1058286426 280 280 280 280
1058286126 266 266 266 266
1058286000 264 264 266 266
1058285700 263 263 263 263
1058285400 394 394 576 576
1058285100 708 708 890 890
1058284800 600 600 890 890
1058284500 345 345 546 546
1058284200 403 403 546 546
1058283900 277 277 376 376
1058283600 303 303 376 376
1058283300 223 223 250 250
1058283000 242 242 250 250
1058282700 227 227 233 233
1058282400 212 212 220 220
1058282100 210 210 220 220
1058281800 212 212 220 220
1058281500 208 208 216 216
1058281200 216 216 216 216
1058280900 203 203 216 216
1058280600 186 186 186 186
1058280300 186 186 186 186
1058280000 198 198 216 216
1058279700 276 276 360 360

Config file :
Target[ARCOpenCase]: `osql -ssqlmonitor -dmonitor -h-2 -E -Q"set
nocount on;select top 1 duration into #t from arc order by [date] desc
;insert into #t select * from #t;select * from #t`
#
# Target returns two identical numbers like this :
# 420
# 420
#
MaxBytes[ARCOpenCase]: 10000
AbsMax[ARCOpenCase]: 10000
Title[ARCOpenCase]: ARCOpenCase
PageTop[ARCOpenCase]:
        <link type="text/css" rel="stylesheet" href="../../sa/graystyle.css">
        <H2 class=heading2>Time to open case in ARC</H2>
Options[ARCOpenCase]: gauge, noinfo
XSize[ARCOpenCase]: 320
YSize[ARCOpenCase]: 100
YLegend[ARCOpenCase]: msec
ShortLegend[ARCOpenCase]: ms
LegendI[ARCOpenCase]: ARCOpenCase:
LegendO[ARCOpenCase]: ARCOpenCase:

Can anyone explain whats going on here?

TIA

 
 
 

1. MRTG storing erroneous gauge data

I am using MRTG 2.9.17 and rateup to collect ufs data from a sun
management center agent.  The filesystem is completely static.  I have
it set as a gauge:

Options[_]: growright,gauge,unknaszero,noo

Everytime I start collecting from scratch, from one run to the next,
my log file changes as follows:

1008687903 990576 990576
1008687903 990576 990576 990576 990576
1008687606 0 0 0 0
...

5 minutes later...
1008688202 990576 990576
1008688202 990576 990576 990576 990576
1008687903 990576 990576 990576 990576
1008687900 970764 970764 990576 990576 <= ???
1008687600 0 0 0 0
...

Where did that 970764 come from?  This is completely incorrect.
990576 is the correct value, and is returned everytime the agent is
polled (verified with a sniffer).  It is as if the value is
interpolated (the time changes by 3 seconds), but that isn't correct
behavior for a gauge.

How do I prevent this from happening?  Is this a bug?

TIA,
Ian

2. Backpack CD Rom Drive ?

3. Storing MRTG data for statistical analysis

4. FS: PPS Mercury 68040 28 Mhz w/ 16 Megabytes

5. mrtg-2.9.27 on FreeBSD-4.6.2 generating faulty log data

6. Examples of What Will Happen During Y2K

7. Store-to-Store Link?

8. Can't send mails containing "."

9. Cisco 2620 - Throughput, Commit Burst, and Excess Burst data stored in the MIB

10. Using modems to store data on CD

11. Store Modulated Data onto tape recorder

12. Store RS232 data onto tape recorder

13. >> Store old data in CD-ROM <<