[collectd] Strange SNMP collection glitches

Mirko Buffoni briareos at eswat.org
Mon Jan 18 10:20:20 CET 2010


Hi friends,

I had problems with SNMP collection of if_octets-WAN of three
different Zyxel SHDSL routers.  I had recurrent peaks, which let
me think it would have been a problem with overflow.
Florian suggested to dump via CSV to get the real values.

This is the dump of CSV in the around of one of those problems:

Time,RX,TX
1263460913,1909718650,4280983380
1263460928,1909818705,4282152739
1263460943,1909957261,4284020628
1263460958,1910070689,4285672398
1263460973,1910121416,8511173
1263460988,1910206467,9548165
1263461003,1910285923,11029693
1263461018,1910367109,13124294
1263461033,1910459026,15791119
1263461048,1910530275,994360
1263461063,1910579982,2860001
1263461078,1910625075,4076156
1263461093,1910653726,4329379
...
1263631988,2244878778,4283169076
1263632003,2244968360,4283484312
1263632018,2245014854,4283679452
1263632033,2246634960,11923400
1263632048,2248276888,1382390
1263632063,2249221976,5028316
1263632078,2249312207,5370610
...
1263682178,2722994326,4286110543
1263682193,2723032355,4286421294
1263682208,2723041634,4286450709
1263682223,2723088914,8842626
1263682238,2723121268,9348159
1263682253,2723161619,10036627
1263682268,2723210595,10939271
1263682283,2723255549,11796160
1263682298,2723307684,12317803
1263682313,2723358266,13210342
1263682328,2723408796,14212347
1263682343,2723430230,14600396
1263682358,2723451638,15018784
1263682373,2723502378,15536030
1263682388,2723550961,15887076
1263682403,2723555542,15896454
1263682418,2723569657,16029926
1263682433,2723575034,16061593
1263682448,2723601784,16497887
1263682463,2723626551,5671
1263682478,2723687005,88056
1263682493,2723771289,371539
1263682508,2723781976,436543
1263682523,2723787778,445971
1263682538,2723805515,1004508

As you can see, there is no router reset (rx is still a valid value),
while TX counter goes nuts for some time after overflowing.


I have three different Zyxel SHDSL routers  (Different firmwares) but
they have this same behavior.  Could it be a bug in the firmware?

For now I solved by fixing a maximum range to the rrd database, so I
have blanks in place of peaks.  I'd be glad to hear if this behavior
can be corrected in some other way or not.

Mirko                   




More information about the collectd mailing list