host down , no more graph

[taihia]

Hi everydoy,

Some of my host are down, it doesn't grah anymore but

when i click on it , it steal give back information by the snmp sysdesc,

when i try to ping from the cacti server it is reachable, and when i do a snmtget from the bash it steal give me the right information.

all the device are in a separate network but with other device that steal graph. (same type of device)

Does some have a idea ?

Thanks you

[gandalf]

As a start, please consult 2nd link of my sig
Reinhard

[taihia]

thank's you for you response.

I starting reading your debuggind information , i try to give more information:

-All the device went down at the same time. only 8 out of 70
-I use CACTI 0.8.6.i
-i use cmd.php
when i turn the graph in debug moke rrdtools is ok (see the file)

what else , i can't find the id of the device as Reinhard wrote in his document ? find the id in the url ?

i will be back if i pogress in the debuging

don't hesitate to post idea

thank you again

[taihia]

Hi here is what i do :

it is not came from the timout response as it is say in the first step of debuging i didn't see anythink like these in the logs

the snmpwalks work very well as i do on the device that is down:it give me all the response of the mibs

i check the poller.php and i have these for all the device that's is down:

ping:default
snmp:snmp not performed due to setting or ping result

i check for the update and i have the update.

I will continue the debugging tomorrow .
anyone have a idea.

thank you reinhard for the doc.

[taihia]

hi everyone,

i have this in the logs concerning one host that is steal down,

03/25/2009 10:35:42 AM - CMDPHP: Poller[0] Host[2] PING: default
03/25/2009 10:35:42 AM - CMDPHP: Poller[0] Host[2] SNMP: SNMP not performed due to setting or ping result.[/i]

i try to increase the ping timout time to 1500ms and the snmp to 500ms but nothing happens.

i am using poller.php with a cacti 0.8.6i.


here is the rrtool info concerning the rra:

[root@supervision rra]# rrdtool info hdp_1_snmp_oid_15.rrd
filename = "hdp_1_snmp_oid_15.rrd"
rrd_version = "0001"
step = 60
last_update = 1237582501
ds[snmp_oid].type = "GAUGE"
ds[snmp_oid].minimal_heartbeat = 120
ds[snmp_oid].min = -1.0000000000e+02
ds[snmp_oid].max = 1.0000000000e+02
ds[snmp_oid].last_ds = "UNKN"
ds[snmp_oid].value = 5.4200000000e+01
ds[snmp_oid].unknown_sec = 0

but the data inside are negative so i want to change the ds.min and ds.max so i do these:

rrdtool tune *.rrd --minimum ds[snmp_oid].min = -1.0000000000e+02:ds[snmp_oid].min = -0.700000000e+02

but here is i have i think it is a problem in the syntax:

ERROR: unknown data source name 'ds'

i also do :

rrdtool fetch the_name_of_rra.rrd AVERAGE and here is what i have:

1238013780: nan
1238013840: nan
1238013900: nan
1238013960: nan
1238014020: nan
1238014080: nan
1238014140: nan
1238014200: nan
1238014260: nan
1238014320: nan
1238014380: nan
1238014440: nan


Did anyone can help me Thank you.

[taihia]

I have somethink new in my problems

i change the downed host detection to snmp reliable only
and it graph again.

But it bring me another question :

why when you use ping and snmp it show the host down ?
becaue when i ping from the server i can ping the host ?

[TheWitness]

Because it's an old cacti version.

TheWitness

[taihia]

thank you for your reply