weird graph
Moderators: Developers, Moderators
weird graph
what can cause such a weird graphs ?
this is the only troubled interfaces among 166 interfaces on that host.
cacti host runs Solaris x86/10, cacti 0.8.6.j, rrdtool 1.2.23.
target host runs Cisco IOS.
Looks like some threshold is overriden, but on that host there are lots of interfaces with much bigger traffic....
You also may think that in the moments of time when there's no data on graphs this host is unreachable - but that's not true, those gaps are specific only to this interface.
Thanks.
this is the only troubled interfaces among 166 interfaces on that host.
cacti host runs Solaris x86/10, cacti 0.8.6.j, rrdtool 1.2.23.
target host runs Cisco IOS.
Looks like some threshold is overriden, but on that host there are lots of interfaces with much bigger traffic....
You also may think that in the moments of time when there's no data on graphs this host is unreachable - but that's not true, those gaps are specific only to this interface.
Thanks.
- Attachments
-
- graph01.png (38.75 KiB) Viewed 2946 times
-
- graph02.png (40.63 KiB) Viewed 2946 times
Typically that type of looking graph would be because the 32bit counters are overrunning (like on a 1Gbit switch port). But since thats only ~60K, hard to believe that is happening.
run rrdtool info file.rrd for the rrd file that graph uses and post the contents.
run rrdtool info file.rrd for the rrd file that graph uses and post the contents.
| Scripts: Monitor processes | RFC1213 MIB | DOCSIS Stats | Dell PowerEdge | Speedfan | APC UPS | DOCSIS CMTS | 3ware | Motorola Canopy |
| Guides: Windows Install | [HOWTO] Debug Windows NTFS permission problems |
| Tools: Windows All-in-one Installer |
- gandalf
- Developer
- Posts: 22383
- Joined: Thu Dec 02, 2004 2:46 am
- Location: Muenster, Germany
- Contact:
I assume that you're rightBSOD2600 wrote:Typically that type of looking graph would be because the 32bit counters are overrunning (like on a 1Gbit switch port). But since thats only ~60K, hard to believe that is happening.
run rrdtool info file.rrd for the rrd file that graph uses and post the contents.
Did you pay attention to the x-axis? It's second consolidation level (week)! And as most users display AVERAGEs only (because cacti does it this way), you will notice consolidated AVERAGEs instead of MAXimum.
To sum up: change to SNMP V2 and use 64 bit counters for those devices
Reinhard
Changed to SNMPv2, that didn't help (do I need to clear poller cache or do some other actions ? I did nothing except simply change the snmp version in target host's properties, and pressed "save").
I've dumped this host RRA and had a look into it. Yeah, all the "gaps" are filled with "NaN" values. :/ Seems quite strange to me.
I did the following:
If I am correct, then those values can be accessed via snmp. So, what's wrong, how can I do further debugging ?
I've dumped this host RRA and had a look into it. Yeah, all the "gaps" are filled with "NaN" values. :/ Seems quite strange to me.
I did the following:
Code: Select all
twilight 16:53:30 snmp_queries#snmpwalk 192.168.3.15 ifDescr | grep Tunnel64
IF-MIB::ifDescr.101 = STRING: Tunnel64
twilight 16:56:55 snmp_queries#snmpwalk 192.168.3.15 .1.3.6.1.2.1.2.2.1.10 | grep .101\ =
IF-MIB::ifInOctets.101 = Counter32: 1945707614
- fmangeant
- Cacti Guru User
- Posts: 2345
- Joined: Fri Sep 19, 2003 8:36 am
- Location: Sophia-Antipolis, France
- Contact:
Hi
could you post your poller statistics ?
Something like :
Maybe you have 2 cron entries, which can make trouble...
could you post your poller statistics ?
Something like :
Code: Select all
$ grep STATS log/cacti.log
[size=84]
[color=green]HOWTOs[/color] :
[list][*][url=http://forums.cacti.net/viewtopic.php?t=15353]Install and configure the Net-SNMP agent for Unix[/url]
[*][url=http://forums.cacti.net/viewtopic.php?t=26151]Install and configure the Net-SNMP agent for Windows[/url]
[*][url=http://forums.cacti.net/viewtopic.php?t=28175]Graph multiple servers using an SNMP proxy[/url][/list]
[color=green]Templates[/color] :
[list][*][url=http://forums.cacti.net/viewtopic.php?t=15412]Multiple CPU usage for Linux[/url]
[*][url=http://forums.cacti.net/viewtopic.php?p=125152]Memory & swap usage for Unix[/url][/list][/size]
[color=green]HOWTOs[/color] :
[list][*][url=http://forums.cacti.net/viewtopic.php?t=15353]Install and configure the Net-SNMP agent for Unix[/url]
[*][url=http://forums.cacti.net/viewtopic.php?t=26151]Install and configure the Net-SNMP agent for Windows[/url]
[*][url=http://forums.cacti.net/viewtopic.php?t=28175]Graph multiple servers using an SNMP proxy[/url][/list]
[color=green]Templates[/color] :
[list][*][url=http://forums.cacti.net/viewtopic.php?t=15412]Multiple CPU usage for Linux[/url]
[*][url=http://forums.cacti.net/viewtopic.php?p=125152]Memory & swap usage for Unix[/url][/list][/size]
- gandalf
- Developer
- Posts: 22383
- Joined: Thu Dec 02, 2004 2:46 am
- Location: Muenster, Germany
- Contact:
That won't help. It's only the first step. You will have to create new interface graphs using the 64bit templates this timedrookie wrote:Changed to SNMPv2, that didn't help (do I need to clear poller cache or do some other actions ? I did nothing except simply change the snmp version in target host's properties, and pressed "save").
Reinhard
okay, thanks for that advice. I recreated that graph using the 64-bits template. But the problem persisted in following: my cisco router doesn't support the ifHCInOctets MIB, so he gives me no data (I've checked in "sh snmp mib" command).
Sorry for the dumbiness, but I just don't understand - what's the difference between 2 MIBs - ifHCInOctets and ifInOctets, since my cisco shows valid data when asked ifInOctets ? The data exceeds the 32 bits range, but seems still valid... Plus, as I said before, I have manyother interfaces those counter have much bigger values.
May there be some other workaround, rather than this MIB unsupported by cisco ?
Sorry for the dumbiness, but I just don't understand - what's the difference between 2 MIBs - ifHCInOctets and ifInOctets, since my cisco shows valid data when asked ifInOctets ? The data exceeds the 32 bits range, but seems still valid... Plus, as I said before, I have manyother interfaces those counter have much bigger values.
May there be some other workaround, rather than this MIB unsupported by cisco ?
Last edited by drookie on Fri Nov 30, 2007 8:47 am, edited 1 time in total.
Is your cisco device one of the ones listed in here ? If yours doesn't support SNMPv2, which would be really odd for a cisco device... the only other option you have is to switch to 1min polling.RFC 2233 wrote: ifHCInOctets OBJECT-TYPE
SYNTAX Counter64
MAX-ACCESS read-only
STATUS current
DESCRIPTION
"The total number of octets received on the interface,
including framing characters. This object is a 64-bit
version of ifInOctets.
Discontinuities in the value of this counter can occur
at re-initialization of the management system, and at
other times as indicated by the value of
ifCounterDiscontinuityTime."
::= { ifXEntry 6 }
| Scripts: Monitor processes | RFC1213 MIB | DOCSIS Stats | Dell PowerEdge | Speedfan | APC UPS | DOCSIS CMTS | 3ware | Motorola Canopy |
| Guides: Windows Install | [HOWTO] Debug Windows NTFS permission problems |
| Tools: Windows All-in-one Installer |
Mhm... I was completely wrong and forgot to use -v 2c key when testing from shell. It really works.
But. I still see a traffic on this graph only when it drops below 32-bit threshold. I recreated the graph using new 64-bit template, but still...
Is it enough to just change the snmp version inside the cacti device's proterties, or do I have to recreate all the graphs and may be the whole device for this graph to start working ?
But. I still see a traffic on this graph only when it drops below 32-bit threshold. I recreated the graph using new 64-bit template, but still...
Is it enough to just change the snmp version inside the cacti device's proterties, or do I have to recreate all the graphs and may be the whole device for this graph to start working ?
Who is online
Users browsing this forum: No registered users and 1 guest