Trouble adding new devices

[tchiverton]

When ever I add a new device to our Cacti install, it fails to be spotted as 'Up' and graphs all fail. I have tracked this down to no data source being created (even when going through the sort-of wizard.
So, i've tried to add a datasource manualy, and the 'data source debug' area is empty !

Any ideas where to start looking for errors ? There is nothing unusaly in the DEBUG level log output.

We are running the latest version, the only oddness I can see is that there are only two things in the Data Templates page, is that normal ?

[tchiverton]

Odly, when I go to create a datasource, all the remplates are listed in the drop down.

Create a data source without a template seems to work ok apart from
Warning: Undefined index: in /var/www/html/cacti-0.8.6e/lib/rrd.php on line 207
/usr/local/rrdtool/bin/rrdtool create \
/var/www/html/cacti-0.8.6e/rra/sadc_server_ds_252.rrd \
--step 300 \
DS:status:GAUGE:600:0:100 \
RRA::0.5:6:700 \
RRA:AVERAGE:0.5:1:600 \
RRA:AVERAGE:0.5:6:700 \
RRA:AVERAGE:0.5:24:775 \
RRA:AVERAGE:0.5:288:797 \
RRA:MAX:0.5:1:600 \
RRA:MAX:0.5:6:700 \
RRA:MAX:0.5:24:775 \
RRA:MAX:0.5:288:797 \

Running that command as cactiuser says:
ERROR: can't parse argument 'RRA::0.5:6:700'


Where/what are the two queries ? Is a join or something breaking somewhere ?

I notice a template import menu item, could I use this to re-import the missing templates ( I am missing some, yes ?)

[tchiverton]

The problem appears to be the
select
data_template.id,
data_template.name,
data_input.name as data_input_method,
data_template_data.active as active
from data_template,data_template_data
left join data_input on data_template_data.data_input_id = data_input.id
where data_template.id = data_template_data.data_template_id
and data_template_data.local_data_id = 0
order by data_template.name
query in data_template.php. If I remove the data_template_data.local_data_id=0 clause, I get a bit long list back as expected, other wise I get a much shorter list (just ones I've created today for debugging, basicaly).

Now, as far as I can tell, this =0 is used to indicate something like the next free 'slot' for that template, as ones that do show up have several entries, for various other integer values of local_data_id, as well as a '0' one.

So, this should be easy to fix, right ? Am I on the right lines (developers !) ?

[tchiverton]

Anyone here who knows how the databse schema is meant to look ?

[TheWitness]

Don't alter the select statements in the PHP code. Please run a database repair. Looks like your database is damaged. Hopefully it's just damaged indexes.

TheWitness

[tchiverton]

Thanks for the reply !

Don't alter the select statements in the PHP code.

I had come to the conclusion MySQL/Cacti/PHP must have had an accident at some point, which may have lead to entries in the table being wrong.
Or something

Please run a database repair.
Looks like your database is damaged. Hopefully it's just damaged indexes.

TheWitness

Where is this repair option ? Is it a Cacti or MySQL thing ?

[TheWitness]

It's a MySQL thing. Check the documentation on how to repair a damaged MyISAM database.

TheWitness

[tchiverton]

It's a MySQL thing. Check the documentation on how to repair a damaged MyISAM database.

TheWitness

OK, I have done this (it was only connection count errors).

Stil the problem occurs (no data sources for new hosts, few things in data templates page), and in the data_template_data table :

Code: Select all

+-----+-----------------------------+---------------+------------------+
| id  | local_data_template_data_id | local_data_id | data_template_id |
+-----+-----------------------------+---------------+------------------+
| 336 |                         334 |           259 |               67 |
| 334 |                           0 |             0 |               67 |
| 337 |                         333 |           260 |               66 |
| 333 |                           0 |             0 |               66 |
| 335 |                         159 |           258 |               65 |
| 332 |                         159 |           257 |               65 |
| 159 |                           0 |             0 |               65 |
| 194 |                           0 |             0 |               64 |
| 319 |                           0 |             0 |               63 |
| 322 |                         319 |           247 |               63 |
| 321 |                         319 |           246 |               63 |
| 308 |                           0 |             0 |               61 |
| 301 |                           0 |             0 |               58 |
+-----+-----------------------------+---------------+------------------+
.
.
.
+-----+-----------------------------+---------------+------------------+
| id  | local_data_template_data_id | local_data_id | data_template_id |
+-----+-----------------------------+---------------+------------------+
| 181 |                          33 |           117 |               33 |
| 144 |                          33 |            80 |               33 |
|  94 |                          33 |            31 |               33 |
|  76 |                          33 |            14 |               33 |
| 161 |                          32 |            97 |               32 |
| 142 |                          32 |            78 |               32 |
|  92 |                          32 |            29 |               32 |
|  74 |                          32 |            12 |               32 |
| 160 |                          31 |            96 |               31 |
| 143 |                          31 |            79 |               31 |
|  93 |                          31 |            30 |               31 |
|  75 |                          31 |            13 |               31 |
+-----+-----------------------------+---------------+------------------+

This is ordered by data_template_id, and as you can see some have no 0 in local_data_id (still).
Is that normal ? The query makes me thing it isn't...

[TheWitness]

Tony,

Can you take this one...

Larry

[rony]

Please dump and email me your database.

What version of cacti?



Yo, Larry, stop trolling the boards on your vacation..

[tchiverton]

Sure, I'd be happy to email it to you, if you let me have your email address
Mine is tom.chiverton@bluefinger.com

It is the most very recent version of cacti. It started of as a .86, and has been upgraded several times since as new Cacti versions come out.

[rony]

My email address is in the linky buttons under every post I make.

[tchiverton]

My email address is in the linky buttons under every post I make.

I feel a bit silly now

Email on way.

[rony]

Wow, well, finally had a free moment to look at you database....

You are missing most of you data templates that are associated with Data queries and graphs on your system.

I would suggest installing 0.8.6e on another system and exporting the working templates and then import those working templates into your corrupted installation.

[tchiverton]

I would suggest installing 0.8.6e on another system and exporting the working templates and then import those working templates into your corrupted installation.

When I try that (file attached), the page refreshes itself and says "Error: XML: Hash version does not exist."