[rtg] Getting a lot of OOR's and Wraps using an Interval of 900 seconds

[Brad Killebrew]

eat shit.


On Fri, 9 Mar 2007, Landon Stewart | Superb Internet Corp. wrote:

> Hello RTG Community,
>
> We want to use 900 second Interval to decrease the amount of data stored
> instead of every 300 seconds.  It takes just about 300 seconds to complete
> one poll anyway (how can I decrease this by the way?).  We also want to use
> 64 bit counters.  The problem I'm experiencing is out of the 20482 targets
> (all targets are interfaces on cisco routers and switches) approximately 390
> of them give an OOR error when rtgpoll runs for the first time.  The OOR
> problem then gets worse with the second run and so on.
>
> First Poll:
> [03/02 13:14:51 Poll round 1 complete.]
>
> [Polls = 19384] [DBInserts = 610] [Wraps = 0] [OutOfRange = 390]
> [No Resp = 0] [SNMP Errs = 0] [Slow = 0] [PollTime = 306.665s]
>
> Second Poll:
> [03/02 13:29:52 Poll round 2 complete.]
>
> [Polls = 38768] [DBInserts = 13605] [Wraps = 71] [OutOfRange = 659]
> [No Resp = 0] [SNMP Errs = 0] [Slow = 0] [PollTime = 307.847s]
>
> Third Poll:
> [03/02 13:44:52 Poll round 3 complete.]
>
> [Polls = 58152] [DBInserts = 26623] [Wraps = 145] [OutOfRange = 927]
> [No Resp = 0] [SNMP Errs = 0] [Slow = 0] [PollTime = 307.625s]
>
> 285th Poll:
> [03/09 14:31:15 Poll round 285 complete.]
>
> [Polls = 5685174] [DBInserts = 3790435] [Wraps = 31537] [OutOfRange =
> 113330]
> [No Resp = 6] [SNMP Errs = 0] [Slow = 0] [PollTime = 306.831s]
>
> (The success rate is awfully low and the Wraps and OOR's are excessive.)
>
> RTG.CONF:
> Interval        900
> HighSkewSlop    3
> LowSkewSlop     .5
> OutOfRange      1152921504606846975
> SNMP_Ver        2
> SNMP_Port       161
> Threads 5
> [snip]
>
> Very Small Snippet of targets.cfg, one showing a seemingly reasonable OOR
> value of and the other 0:
> 10.36.225.4      .1.3.6.1.2.1.31.1.1.1.6.134     64      public
> ifInOctets_1    108     108000000000     (3/48)
> 10.36.225.4     .1.3.6.1.2.1.31.1.1.1.7.134     64      public
> ifInUcastPkts_1 108     108000000000     (3/48)
> 10.36.225.4     .1.3.6.1.2.1.31.1.1.1.10.134    64      public
> ifOutOctets_1   108     108000000000     (3/48)
> 10.36.225.4     .1.3.6.1.2.1.31.1.1.1.11.134    64      public
> ifOutUcastPkts_1        108     108000000000     (3/48)
> 10.36.225.4     .1.3.6.1.2.1.31.1.1.1.6.139     64      public
> ifInOctets_1    109     0        (VLAN-701)
> 10.36.225.4     .1.3.6.1.2.1.31.1.1.1.7.139     64      public
> ifInUcastPkts_1 109     0        (VLAN-701)
> 10.36.225.4     .1.3.6.1.2.1.31.1.1.1.10.139    64      public
> ifOutOctets_1   109     0        (VLAN-701)
> 10.36.225.4     .1.3.6.1.2.1.31.1.1.1.11.139    64      public
> ifOutUcastPkts_1        109     0        (VLAN-701)
>
> Small Snippet of rtgpoll output with -v:
> *** Out of Range (10.160.63.2 at .1.3.6.1.2.1.31.1.1.1.7.149) [insert_val:
> 18446743581360759826] [oor: 0]
> *** Out of Range (10.160.63.3 @.1.3.6.1.2.1.31.1.1.1.7.149) [insert_val:
> 1163086] [oor: 0]
> *** Counter Wrap (10.160.63.2 at .1.3.6.1.2.1.31.1.1.1.11.144) [poll:
> 879969481][last: 112452383790][insert: 18446743962137137306]
> *** Out of Range (10.160.63.2 @.1.3.6.1.2.1.31.1.1.1.11.144) [insert_val:
> 18446743962137137306] [oor: 0]
> *** Out of Range (10.160.63.3 at .1.3.6.1.2.1.31.1.1.1.11.144) [insert_val:
> 6900311] [oor: 0]
> *** Out of Range (10.160.63.2 at .1.3.6.1.2.1.31.1.1.1.7.144) [insert_val:
> 70775968] [oor: 0]
> *** Out of Range (10.160.63.2 at .1.3.6.1.2.1.31.1.1.1.10.160) [insert_val:
> 1088853] [oor: 0]
> *** Out of Range (10.36.225.4 at .1.3.6.1.2.1.31.1.1.1.11.26) [insert_val:
> 11163020] [oor: 0]
> *** Out of Range (10.36.225.4 @.1.3.6.1.2.1.31.1.1.1.11.141) [insert_val:
> 33300] [oor: 0]
> *** Out of Range (10.36.225.4 at .1.3.6.1.2.1.31.1.1.1.6.7) [insert_val:
> 16490194] [oor: 0]
>
> What am I doing wrong?  What settings to I need to pay attention to here and
> is the rtgtargetmkr.pl script working properly when it sets the OOR to 0 for
> some interfaces?  How do I calculate the correct values for the rtg.conf and
> why doesn't the rtgtargetmkr.pl script use the value from rtg.conf for the
> OOR values it can't fill in?
>
> (IP addresses have been obfuscated.)
>
> -- 
> Landon Stewart <LStewart at SUPERB.NET>
> SuperbHosting.Net by Superb Internet Corp.
> Toll Free: 888-354-6128 x 4199 (US/Canada)
> International: 604-638-2525 x 4199
> Web hosting and more "Ahead of the Rest": http://www.superbhosting.net
>