poller.php is not updating rrd files

[dknell]

I have read ALL of the posts in here and none of the solutions seem to fix this issue... Nothing has changed on my server except MySQL died because the log partition filled up. I moved the MySQL logs, and brought it back up and now Cacti does not update the rrd files. I tried restarting Cacti and MySQL, clearing the poller cache, changing perms on rra and log dirs to no avail.

Running:
SuSE 8.2
PHP 4.3.8
Cactid 0.8.6c
rrdtool 1.0.49

The permissions of the rra and log dirs are 777.

The rrd files haven't been touched since MySQL died. I have tried to run poller.php manually as the correct cacti user and there are no errors that indicate what the problem may be. I will post the logs of when I ran poller.php manually.

Please help!

[rony]

Your poller cache is probably hosed. The solution is the clear the poller cache. This can be done from the System Utilities menu.

Let me know if this resolves your issue.

[dknell]

I have tried clearing the cache several times to no avail. Is there a way I can clear the cache from command line?

[dlaperle]

I have the same problem...

Clearing the cache don't seem to fix the problem...

Maybe something else, it's hard to define where is the problem...

I'll be glad to heard an answer from some BIG BOSSES of cacti

[rony]

Run the following querys on the cacti database:

Code: Select all

delete from poller_item;

Then run:

Code: Select all

select count(*) as total from poller_item;

If the return result is greater than 0 there is a problem with the table and I would suggest running the following to fix it:

Code: Select all

DROP TABLE poller_item;

CREATE TABLE `poller_item` (
  `local_data_id` mediumint(8) unsigned NOT NULL default '0',
  `poller_id` smallint(5) unsigned NOT NULL default '0',
  `host_id` mediumint(8) NOT NULL default '0',
  `action` tinyint(2) unsigned NOT NULL default '1',
  `hostname` varchar(250) NOT NULL default '',
  `snmp_community` varchar(100) NOT NULL default '',
  `snmp_version` tinyint(1) unsigned NOT NULL default '0',
  `snmp_username` varchar(50) NOT NULL default '',
  `snmp_password` varchar(50) NOT NULL default '',
  `snmp_port` mediumint(5) unsigned NOT NULL default '161',
  `snmp_timeout` mediumint(8) unsigned NOT NULL default '0',
  `rrd_name` varchar(19) NOT NULL default '',
  `rrd_path` varchar(255) NOT NULL default '',
  `rrd_num` tinyint(2) unsigned NOT NULL default '0',
  `arg1` varchar(255) default NULL,
  `arg2` varchar(255) default NULL,
  `arg3` varchar(255) default NULL,
  PRIMARY KEY  (`local_data_id`,`rrd_name`),
  KEY `local_data_id` (`local_data_id`),
  KEY `host_id` (`host_id`)
) TYPE=MyISAM;

Then with either case, run the Clear Poller Cache again from the System utilities.

The Clear Poller Cache is suppose to do this, but I suspect that your table is corrupted due to the other issue you reported. Droping the table and recreating it should fix that. Please remember to stop your cron/scheduler so that the poller does not attempt to run when you are working on the table.

Let me know what you find.

[dlaperle]

I just try what you said...

i did: delete from...

after the select...and it return 0 ... so i didnt make the drop and the create database stuff...

I'm so lost now....i really don't want to reinstall i really don't want to lost all my data gathered so far...

Can i reinstall and reput all the data i have or something?

Thks a lot for the help!

[dknell]

I just tried what you suggested, and it is still not updating.

mysql> delete from poller_item;
Query OK, 3133 rows affected (0.01 sec)

mysql> select count(*) as total from poller_item;
+-------+
| total |
+-------+
| 0 |
+-------+
1 row in set (0.00 sec)

Then I cleared the poller cache from System Utilities and then checked the poller_item table again and it was re-populated....

mysql> select count(*) as total from poller_item;
+-------+
| total |
+-------+
| 3133 |
+-------+
1 row in set (0.00 sec)

Any more suggestions?

Thanks again for the quick response...

[rony]

Well, um, at this point, I want you to post your cacti log from one poll cycle with the logging setting set to DEBUG. That setting is located in Settings -> General -> Poller Logging Level. You can find where the log file is located by checking the Settings -> Paths -> Cacti Log File Path

If the log is to long, please email it to me rony at disorder dot com.

[rony]

Sorry I'm tired today, saw the attachment in the first page.

First, stop the poller from running and execute the following command:

Code: Select all

select count(*) as total from poller_output;

If the total is greater than 0, run the following query:

Code: Select all

delete from poller_output;

Let me know the results.

[Guest]

WOOHOO!!!!!!!!!!!!!!

You are the MAN!!!!!!

That did it... Thanks so much for your help!!!!