Issues with SNMP Version 3

[bawayce]

Long time user first time poster. I have used the tips and troubleshooting from these forms for two years and find them very valuable. I believe I have done my due diligence in attempting to resolve this issue without having to post but to no avail so here I am.

Setup:
2008 R2 64 bit
IIS 7.5
Cacti 8.8 and supported apps loaded from Cacti Windows Installer
Some patching to get past some of the 8.8 bugs

Environment:
I have a working Cacti install on UNIX with 8.6 graphing 100+ nodes and 1000's of graphs but the system is nearing it's expiration date and our IT director want this system ported to a Windows environment since this is our only UNIX system in production. This system is currently using SNMP version 3 to poll all of our monitored devices without issues.

So new system running 8.8 can graph and poll without issues if it's using SNMP Version 2. Once I set it to use Version 3 the systems go from alive to dead when using any of the SNMP methods for Downed Device Detection. If I set the detection to ICMP the device show as alive but the poller does not report any data.

Logs with the ICMP device detection show:

Code: Select all

05/10/2012 03:50:02 PM - SPINE: Poller[0] Host[3] DEBUG: ICMP Host Alive, Try Count:1, Time:16.0000 ms
05/10/2012 03:50:02 PM - SPINE: Poller[0] Host[3] PING: Result ICMP: Host is Alive
05/10/2012 03:50:02 PM - SPINE: Poller[0] Host[3] TH[1] RECACHE: Processing 1 items in the auto reindex cache for 'da01_sw01.xxx.xxx'
05/10/2012 03:50:02 PM - SPINE: Poller[0] Host[3] TH[1] Recache DataQuery[1] OID: .1.3.6.1.2.1.1.3.0, output: U

I did some research on the output:U but could not isolate anything specific to SNMP V3 or 8.8.

Clients are Cisco devices running lastest IOS 15.1

I have verified and re-verified that the SNMP V3 setting are correct.

Any idea?

[BSOD2600]

Unfortunately, I don't have any snmpv3 enabled devices to troubleshoot with.

1) if you switch to cmd.php, does snmpv3 queries work?
2) net-snmp components are installed?
3) A snmpwalk v3 walk from the commandline on the Cacti server against a problem device works?
4) you're aware cacti 0.8.8a is out?

[bawayce]

Hey BSOD thanks for the reply.

The necessary SNMP components (crypto and openssl) are installed.

An smnpwalk using version 3 is successful

Code: Select all

snmpwalk -v3 -u cactiuser -l AuthPriv -a MD5 -A xxxxxxx -x DES
 -X xxxxxxx 10.123.xxx.xxx .1.3.6.1.2.1.1.3.0
DISMAN-EVENT-MIB::sysUpTimeInstance = Timeticks: (237728419) 27 days, 12:21:24.19

I upgraded to 8.8a. Still no improvements.

HOWEVER I retried moving the poller to CMD and that did work. I know I tried this in the past and it did not make a difference. Not sure what the change was. Upgrade?

So any ideas as to why spine is having issues?

[BSOD2600]

So any ideas as to why spine is having issues?

Running spine 0.8.8a too?
Tried running it in manually in verbose mode for one of those snmpv3 devices to see if it gives any more useful information?

Otherwise TheWitness is the primary expert on the tool... and from his FB page, it would appear he's enjoying a tropical vacation, so not going to be much help for a bit.