No more graph display

[ewok2]

Hello
I have a cacti Version 1.1.38 on a Ubuntu
Some days ago I have lost some curve on some device. Nothing is display till this morning.

When I am logged on the Ubuntu, a snmpwalk answer correctly.
On the device page of cacti, the Host is considered Up (with snmp uptime).

The rra files are updated (at least the date of the file is changed)

But the Cacti log send this message :

Code: Select all

2018/12/10 15:25:13 - PCOMMAND Device[4] Device[rasp-xx] RECACHE: Recache successful.
AND enabled='on' - found: 0
WHERE snmp_query_id=9
FROM automation_graph_rules AS agr
agr.snmp_query_id, agr.graph_type_id
2018/12/10 15:25:13 - AUTOM8 TRACE automation_execute_data_query Device[4] Device[rasp-xx] - sql: SELECT agr.id, agr.name,
2018/12/10 15:25:13 - AUTOM8 TRACE automation_execute_data_query Device[4] Device[rasp-xx] - start - data query: 9
2018/12/10 15:25:12 - PCOMMAND Device[4] Device[rasp-xx] RECACHE: Recache for Device, data query #9
2018/12/10 15:25:12 - PCOMMAND Device[4] Device[rasp-xx] RECACHE: Recache successful.
AND enabled='on' - found: 0
WHERE snmp_query_id=2
FROM automation_graph_rules AS agr
agr.snmp_query_id, agr.graph_type_id
2018/12/10 15:25:12 - AUTOM8 TRACE automation_execute_data_query Device[4] Device[rasp-xx] - sql: SELECT agr.id, agr.name,
2018/12/10 15:25:12 - AUTOM8 TRACE automation_execute_data_query Device[4] Device[rasp-xx] - start - data query: 2
2018/12/10 15:25:12 - PCOMMAND Device[4] Device[rasp-xx] RECACHE: Recache for Device, data query #2
2018/12/10 15:25:12 - PCOMMAND Device[4] Device[rasp-xx] RECACHE: Recache successful.
AND enabled='on' - found: 0
WHERE snmp_query_id=1
FROM automation_graph_rules AS agr
agr.snmp_query_id, agr.graph_type_id
2018/12/10 15:25:12 - AUTOM8 TRACE automation_execute_data_query Device[4] Device[rasp-xx] - sql: SELECT agr.id, agr.name,
2018/12/10 15:25:12 - AUTOM8 TRACE automation_execute_data_query Device[4] Device[rasp-xx] - start - data query: 1
2018/12/10 15:25:12 - PCOMMAND Device[4] Device[rasp-xx] RECACHE: Recache for Device, data query #1
2018/12/10 15:25:12 - PCOMMAND Device[4] Device[rasp-xx] WARNING: Recache Event Detected for Device 

This message occur every 5 minutes.

Any idee of the problem and how to solve it ?

[netniV]

That message is likely unrelated. It is a debug message (not sure why you have it in debug) that is showing that automation found no new devices/graphs to create. Your automation network is probably set to scan every five minutes.

[ewok2]

you mean that the Warning at the end

Code: Select all

2018/12/10 15:25:12 - PCOMMAND Device[4] Device[rasp-xx] WARNING: Recache Event Detected for Device 

Is not a problem ? I thought that Warning mean problem...

Where can I tune the automation network scan period to see if acting on the message ?

[ewok2]

I have tried to execute the cmd.php by hand and in this case the rra get the values and the graph get one point.
But at the next crontab activation it still does not work.

Any idea where to find more information on log ?

[netniV]

When you executed by hand, did you use "sudo -u <user>" at the front to make sure it was running as the same user that cron uses?

[ewok2]

No I just go in the cacti/site directory.
And send

Code: Select all

sudo php cmd.php --first=<ID> --last=<ID>

If I add www-data with -u after the sudo I get the same result, it works.

And a new thing... after these manual action with cmd (with or witout www-data parameter) it seem's to continue to work ????
As if this manual commande repaire some things...

[ewok2]

I was to confident... after 5 minutes more it does not work any more ...

SO what is the difference between the command execute manually and the crontab action ???

[netniV]

It would suggest that your cron entry is wrong. If you are using crontab, you must specify the user you are editing on the command line so there should not be a user specified in your entry.

The reason it probably works for 5 minutes is that you have specified your cron interval within cacti to be 5 minutes, so cmd.php keeps running for up to 5 minutes issuing commands as it needs to.

[ewok2]

Thanks for reply
But I have 18 host with graph associated.
Some Graph works and other not.

Il fact it seem's I have 3 categories of device/graph

The one that works everytime
The one that stop working each time the cacti server or the device server is reboot and start again working after a certain time (from some hour up to somes days...)
The one that does not work at all!

For each of theses,
- the device are up with snmp uptime
- the cmd.php execute manualy get some data and trace the curve with "one point" for each execution of the cmd.php
- snmpwalk from the cacti server works fine

A am a littel lost and does not to know where to search...