(VERY WEIRD) GOING FROM STATUS UP TO STATUS ERROR

[bengelly]

Hi,

I'm monitoring approx 65 devices using cacti and mactrack... this morning, after the WE, for no known reasons, 16 of my devices' status went to ERROR !!! Just like that !!!

Under the CLI, here is what i get for one of them for eg:

Code: Select all

pc88:/usr/share/cacti/site/plugins/mactrack# php -q mactrack_scanner.php -id=43 -d
DEBUG: HOST: 10.57.240.53 is alive, processing has begun.
DEBUG: WARNING: SITE: IUT, IP: 10.57.240.53, TYPE: U, Device Type Not Found in Device Type Table.
DEBUG: SQL: UPDATE mac_track_devices SET ports_total='1',device_type_id='51',scan_type = '1',vlans_total='0',ports_active='0',ports_trunk='0',macs_active='1',snmp_readstring='',snmp_sysName='U',snmp_sysLocation='U',snmp_sysContact='U',snmp_sysObjectID='U',snmp_sysDescr='U',snmp_sysUptime='U',snmp_status='4',last_runmessage='WARNING: Device Type Not Found in Device Type Table.',last_rundate='2007-02-12 10:04:27',last_runduration='0.0109' WHERE device_id ='43'

What the hell is type "U" ? That U is in all the fields of my mactrack_devices db

The weirdest thing is that these devices, some of them switchs, some of them APs, were perfectly recognised by mactrack after I had manually edited a new device type for them...

I have been running a series of tests, all my devices are alive and return informations to snmp requests, and I can't seem to find what the problem is...

Can someone help please ?

Thanks

[bengelly]

Hi,

the story continues... this afternoon, after launch break, all the troubled devices reported "up" to Mactrack... Everything is green for now...

I changed nothing to my config, I only checked the description and caracteristics of every device in order to be sure that everything was correct and all of the sudden, the correct SNMP info showed up...

What's happening ?!?

Thanks

[TheWitness]

Gremlins.

But seriously, did you change any device types? That would do it. Also, if the "network" went down, the devices rotate the snmp community strings. So, maybe you were using the wrong one at the time. This would indicate that there was a layer2/3 disruption during polling.

TheWitness

[bengelly]

Hi,

sorry the withness but I didn't change any of the device types and the network didn't go down... all the devices in question were responding to SNMP requests under the CLI with the right SNMP info...

as for the SNMP readstrings, I only use one so no possible rotations...

Really weird isn't it ?!?

Thanks

[TheWitness]

It could have also been a timeout issue. Excessive latency will make them appear down. Bottom line is that "something" happened, you just didn't see it and have no record of it.

TheWitness