After creating a graph, an existing graph stopped collecting

[gplante]

I created a graph of the temperature sensed by a module with multiple sensors.
The graph has started collecting data, see picture below.

Graph_MCH1.jpg (77.49 KiB) Viewed 1341 times

In the poller cache there is the following entry:

Pooler_MCH1.jpg (27.01 KiB) Viewed 1341 times

In the hidden snmp, there's this graph:
Details Host: MCH1 , SNMP Query: NetworkPower
Index: 257 , Field Name: customindex , Field Value: 257
OID: Host: MCH1 , SNMP Query: NetworkPower Index: 258 , Field Name: customindex , Field Value: 258
OID: Host: MCH1 , SNMP Query: NetworkPower Index: 513 , Field Name: customindex , Field Value: 513
OID: Host: MCH1 , SNMP Query: NetworkPower Index: 514 , Field Name: customindex , Field Value: 514

OID: Host: MCH1 , SNMP Query: NetworkPower Index: 257 , Field Name: NomModule , Field Value: "PCH1 "
OID: Host: MCH1 , SNMP Query: NetworkPower Index: 258 , Field Name: NomModule , Field Value: "PCH1 "
OID: Host: MCH1 , SNMP Query: NetworkPower Index: 513 , Field Name: NomModule , Field Value: "PCH2 "
OID: Host: MCH1 , SNMP Query: NetworkPower Index: 514 , Field Name: NomModule , Field Value: "PCH2 "

OID: Host: MCH1 , SNMP Query: NetworkPower Index: 257 , Field Name: NomSenseur , Field Value: "External Sensor1"
OID: Host: MCH1 , SNMP Query: NetworkPower Index: 258 , Field Name: NomSenseur , Field Value: "External Sensor2"
OID: Host: MCH1 , SNMP Query: NetworkPower Index: 513 , Field Name: NomSenseur , Field Value: "External Sensor1"
OID: Host: MCH1 , SNMP Query: NetworkPower Index: 514 , Field Name: NomSenseur , Field Value: "External Sensor2"

OID: Host: MCH1 , SNMP Query: NetworkPower Index: 257 , Field Name: StatusSenseur , Field Value: Actif
OID: Host: MCH1 , SNMP Query: NetworkPower Index: 258 , Field Name: StatusSenseur , Field Value: Inactif
OID: Host: MCH1 , SNMP Query: NetworkPower Index: 513 , Field Name: StatusSenseur , Field Value: Inactif
OID: Host: MCH1 , SNMP Query: NetworkPower Index: 514 , Field Name: StatusSenseur , Field Value: Inactif

Later, I created a new graph of the temperature. The graph began to collect data (see images below).
But the graph that I created above has stopped collecting the data (see graph above).

Graph_MCP1.jpg (81.29 KiB) Viewed 1341 times

In the poller cache there is the following entry:

Pooler_MCP1.jpg (29.08 KiB) Viewed 1341 times

In the snmp cache was for this graph: the graph is the same as before.

I created a third graph, there was no colect data. RRD file was not created.
In the poller cache there is the following entry:

Pooler_MFE1.jpg (29.35 KiB) Viewed 1341 times

The RRD files at 16:00 :
mch1_temperature_1_1.rrd modify : 2012-01-31 11:06
mcp1_temperature_1_1.rrd modify : 2012-01-31 16:00

NO RRD file for MFE1.
Server environment:

CACTI 0.8.7e with plugin architectur
Plugin Monitor
PHP 5.2.13
MYSQL 5.1.44
RRDTOOL 1.2.15
CYGWIN 1.5.20

Windows 2003 SP2

Why, this situation and how to repair this?


Gislain Plante
CSST

[gplante]

More information:

Log with option DEBUG:

MCP1

Log_MCP1.jpg (141.04 KiB) Viewed 1335 times

MCH1

Log_MCH1.jpg (79.59 KiB) Viewed 1335 times

For MFE1 no entries in LOG.

As we can see for MCH1, the ss_network script is not running and it's situation started after creating the graph for MCP1.

[gplante]

Hi!

Is there a maximum number of graph that can handle by Cacti?

All new graph does not work, the rrd file (s) are not created.

More information : Technical support:

Technical Support.jpg (230.79 KiB) Viewed 1306 times

MySQL Tables :

Table for poller.jpg (50.87 KiB) Viewed 1306 times

I rebuild the poller cache and it did not change the situation. I still have the problem.

Ghislain Plante
CSST

[gplante]

Hi,

This is not only the temperature graph that stopped working. It's 30 graph that no longer works.

Gislain Plante
CSST

[gandalf]

Hi!

Is there a maximum number of graph that can handle by Cacti?

The largest know installations has > 1,000,000 data sources. You do not seem to hit that limit
R.

[gandalf]

Hi,

This is not only the temperature graph that stopped working. It's 30 graph that no longer works.

Gislain Plante
CSST

It is correct to first look at the Poller Cache to find alll commands that should be executed by the poller.
In case ean entry is missing here, we should look at data source creation (when "Create Graphs for this Host" is used).
Then, using 2nd link of my sig, it helps to study cacti.log as you did already.
If the polling entry is available, but result is invalid, this should be tackled next.

In case polling interval is exceeded, we should look at invalid script timeouts (again, 2nd link of my sig should help).
R.

[gplante]

Hi,

I test cmd.php:

php -q ..../cmd.php 268 268 268 is the ID of Host.

The log with option debug:

logDebug cmd.jpg (101.4 KiB) Viewed 1259 times

We can see in the log before that there is no writing RRD File (CATI2RRD).

In the poller cache items:

pollerCacheItems.jpg (34.05 KiB) Viewed 1259 times

I test SNMP with MIB Browser:

testsnmp.jpg (67.96 KiB) Viewed 1259 times

After the scheduleted poller (poller.php),i have in log:

02/15/2012 02:55:09 PM - CMDPHP: Poller[0] Host[268] RECACHE: Processing 1 items in the auto reindex cache for 'mry1.intra.local.csst.qc.ca'.
02/15/2012 02:55:09 PM - PHPSVR: Poller[0] DEBUG: INC: 'D:\Appls\cacti-0.8.7\scripts\ss_NetworkPower.php' FUNC: 'ss_networkpower'
PARMS: 'mry1.intra.local.csst.qc.ca lirE;resO get Temperature 257'

WHY?


Question:

What is the difference between fping and TPING. Which one is better?

We use TPING Obtain for the latency.

If I use spine not cmd.php, that he will correct my problem?

Thank You

Ghislain Plante

[gandalf]

After the scheduleted poller (poller.php),i have in log:

02/15/2012 02:55:09 PM - CMDPHP: Poller[0] Host[268] RECACHE: Processing 1 items in the auto reindex cache for 'mry1.intra.local.csst.qc.ca'.
02/15/2012 02:55:09 PM - PHPSVR: Poller[0] DEBUG: INC: 'D:\Appls\cacti-0.8.7\scripts\ss_NetworkPower.php' FUNC: 'ss_networkpower'
PARMS: 'mry1.intra.local.csst.qc.ca lirE;resO get Temperature 257'

That's correct. It will run a query against your self-created index which is 257. But the value you received is 238 and we should find an rrdtool update for that value. Which "log verbosity" did you use to get the logs attached? Perhaps it was too low and thus, the required update was not displayed. See again 2nd linkk of my sig for details on logging

If I use spine not cmd.php, that he will correct my problem?

This shouldn't make any difference. You may try both. Sometimes, you find better logging with spine ...
R.

[gplante]

Hi,
I find why some graph were not performed.
I have a process that does not end.
The processe is waiting on a snmpget (Net-snmp 5.4.2.0-1).
Process in wait:

process.jpg (34.96 KiB) Viewed 1225 times

The corresponding commands to process above:
"D:\Appls\php\php-win.exe" -q "d:/appls/cacti-0.8.7/cmd.php" 254 298
cmd.exe /c D:\Appls\php\php-win.exe -q D:/Appls/cacti-0.8.7/script_server.php cmd
D:\Appls\php\php-win.exe -q D:/Appls/cacti-0.8.7/script_server.php cmd
cmd.exe /c D:\Appls\NETSNMP\bin\snmpget.exe -v2c -c XXXXXXXX mry1.intra.local.csst.qc.ca .1.3.6.1.4.1.4779.1.6.1.1.5.1.
D:\Appls\NETSNMP\bin\snmpget.exe -v2c -c XXXXXXXX mry1.intra.local.csst.qc.ca .1.3.6.1.4.1.4779.1.6.1.1.5.1.1 (in wait)

I did manually in DOS :
D:\Appls\NETSNMP\bin\snmpget.exe -v2c -c XXXXXXXX mry1.intra.local.csst.qc.ca .1.3.6.1.4.1.4779.1.6.1.1.5.1.1

result (is very fast):
Cannot find module (NOTIFICATION-LOG-MIB): At line 1 in (none)
Cannot find module (NET-SNMP-AGENT-MIB): At line 1 in (none)
Cannot find module (DISMAN-EVENT-MIB): At line 1 in (none)
SNMPv2-SMI::enterprises.4779.1.6.1.1.5.1.1 = INTEGER: 238

I did the command "cmd.php" for the device that was waiting in the process: (MRY1 device id = 268) at an interval of 5 minutes: php -q D:\Appls\cacti-0.8.7\cmd.php 268 268 (3 times).

The RRD file was updated at every cmd.php:
The log:

MRY1.jpg (136.53 KiB) Viewed 1225 times

The wait occurs only in a process triggered by the poller. Why?

Would it be better to use snmp of php instead of snmp NET-SNMP?

[gandalf]

The wait occurs only in a process triggered by the poller. Why?

Would it be better to use snmp of php instead of snmp NET-SNMP?

I can't tell. This may depend on your OS and I'm not an expert here. Ask in the windows forum, please
R.

[gplante]

Hello,

The problem is corrected.

I installed a new version of NET-SNMP.

The upgrade the version 5.4.2 to version 5.6.1.1.

Thank You for our help.

Ghislain Plante