[SOLVED] - rebooting server = mrtg/snmp graph spikes

Rammoth · n00b Joined: 05 Jan 2005 Posts: 36

Whenever I reboot my gentoo linux server which is always up2date (emerge --sync ; emerge -uD world), after reboot, I am greeted with these outragous max figure spikes, can anyone tell me what is causing this? It royally pisses me off because after a new kernel recompile or some other reason to reboot, my entire snmp graphs get royally rooted due to these outragous spikes. I've even disconnected the network cards physically and rebooted it, every so often on reboot, these massive spikes occur, and it occurs on both interfaces, not just one. I'll get a 14mB/sec spike on both interfaces for no reason!!

What irritates me, is I've got 2 Fedora Core 2 boxes exactly identical configuration (hardware and snmp/mrtg) and I get no such issues. It leads me to believe possibly snmp is screwing itself as it shuts down and vixie cron happens to probe it with mrtg and bam, crap figures, however, this is just my guess.

I've googled/searched this forum to no avail.

Can anyone help!?

Config is as follows, mrtg is executed at a 5 minute interval by vixie-cron with these options

net-analyzer/net-snmp
Latest version available: 5.2.1-r1
Latest version installed: 5.2.1-r1

NIC Modules:
e100 (eth0)
tg3 (eth1)

snmp config:
com2sec local 127.0.0.0/8 public

group MyROGroup v1 local
group MyROGroup v2c local
group MyROGroup usm local

view all included .1 80

access MyROGroup "" any noauth exact all none none

mrtg.conf:

WorkDir: /var/www/bluefusion/admin/mrtg

Options[_]: growright

EnableIPv6: no

#### LAN Interface ####
Target[lan]: 2:public@127.0.0.1:
SetEnv[lan]: MRTG_INT_IP="192.168.115.1" MRTG_INT_DESCR="eth0"
MaxBytes[lan]: 125000000
Options[lan]: growright nopercent nobanner nolegend
Title[lan]: Traffic Analysis for LAN Interface
PageTop[lan]: <H1>Traffic Analysis for LAN Interface</H1>

#### Internet Interface ####
Target[inet]: 3:public@127.0.0.1:
SetEnv[inet]: MRTG_INT_IP="" MRTG_INT_DESCR="eth1"
MaxBytes[inet]: 12500000
Options[inet]: growright nopercent nobanner nolegend
Title[inet]: Traffic Analysis for Internet Interface
PageTop[inet]: <H1>Traffic Analysis for Internet Interface</H1>

Clicky showing outragous spike.

Rammoth · n00b Joined: 05 Jan 2005 Posts: 36

Anyone

Rammoth · n00b Joined: 05 Jan 2005 Posts: 36

*Bump*

Rammoth · n00b Joined: 05 Jan 2005 Posts: 36

*Nudge*

rum · Posted: Thu Jul 07, 2005 7:29 pm Post subject:

Maybe there is a reason.
Setup a script that monitors the output and do a ps -ef or netstat -nap and print the output somewhere.
_________________
Bad spellers of the world, untie!

Rammoth · n00b Joined: 05 Jan 2005 Posts: 36

I've tried and tried and googled and banged my head.

I can find no good reason that when I reboot the box it spikes to 100% usage like this.

Anyone :?:

Rammoth · n00b Joined: 05 Jan 2005 Posts: 36

I've enabled logging.

Before reboot/spike
2005-10-01 01:56:47 -- --snpo: SNMPGet from public@127.0.0.1: -- ifInOctets.2,ifOutOctets.2,sysUptime,sysName
2005-10-01 01:56:47 -- --snpo: SNMPfound -- '1080456', '4432131', '0:02:52', 'server'
2005-10-01 01:56:47 -- --snpo: run snmpget from ifInOctets&ifOutOctets:public@127.0.0.1

After reboot (now logging bad spike)
2005-10-01 02:00:02 -- --snpo: SNMPGet from public@127.0.0.1: -- ifInOctets.2,ifOutOctets.2,sysUptime,sysName
2005-10-01 02:00:02 -- --snpo: SNMPfound -- '567301', '1249648', '0:01:38', 'server'
2005-10-01 02:00:02 -- --snpo: run snmpget from ifInOctets&ifOutOctets:public@127.0.0.1

2005-10-01 02:05:02 -- --snpo: SNMPGet from public@127.0.0.1: -- ifInOctets.2,ifOutOctets.2,sysUptime,sysName
2005-10-01 02:05:02 -- --snpo: SNMPfound -- '985012', '2615845', '0:06:37', 'server'
2005-10-01 02:05:02 -- --snpo: run snmpget from ifInOctets&ifOutOctets:public@127.0.0.1

2005-10-01 02:10:01 -- --snpo: SNMPGet from public@127.0.0.1: -- ifInOctets.2,ifOutOctets.2,sysUptime,sysName
2005-10-01 02:10:01 -- --snpo: SNMPfound -- '1141266', '3022079', '0:11:36', 'server'
2005-10-01 02:10:01 -- --snpo: run snmpget from ifInOctets&ifOutOctets:public@127.0.0.1

Surely someone knows why it's spiking like this?! Help!

Rammoth · n00b Joined: 05 Jan 2005 Posts: 36

Anyone? Please?! I've been trying to solve this mystery for ... forever!

kargig · n00b Joined: 21 Sep 2004 Posts: 27

paste us the part of the .log file that creates the spike...
check the time you get the spike...and extract that part from the log file and paste it here....

I'm sure the explanation is right there...

btw...read this: http://lists.ee.ethz.ch/mrtg/msg23779.html

it describes your problem...

Rammoth · n00b Joined: 05 Jan 2005 Posts: 36

kargig · n00b Joined: 21 Sep 2004 Posts: 27

http://people.ee.ethz.ch/~oetiker/webtools/rrdtool/pub/contrib/removespikes.tar.gz

use the force young luke