Differents graphs on 2 CACTI versions on the same interface

[combellm]

Hi,
I made an upgrade recently, I keep my old server so I can make everythink I want without stop the production server.

My issue is that I watched a graph OF THE SAME INTERFACE OF THE SAME EQUIPMENT on both server and here are the diferent screenshots I took:
On the old one:

On the new one:


As you can see, this is tottaly different, except sometimes where we could guess that this is maybe the opposite.
Anyway this is weird and I checked this is exactly the same configuration, same template and so...

Version 0.8.8a for the new one
version 0.8.7b for the old one

Best regards,

[paulgevers]

Can you be more precise in what you see and what you didn't expect?. I see two graphs showing the same information on two different machines, thus with different content.

What is your issue here?

[combellm]

Can you be more precise in what you see and what you didn't expect?. I see two graphs showing the same information on two different machines, thus with different content.

What is your issue here?

Hi,
My issue is that the two graphs are on two servers. The first one on an old CACTI and the second on a new one.

The graphs are about the same monitored equipment (MARM01) on the same interface (po3).
But I don't have the same data! It should!

Thanks for your interest.

[combellm]

Hi,
I just notice that whe I check the parameters I have:

Code: Select all

	

/usr/bin/rrdtool graph - \
--imgformat=PNG \
--start=-2678400 \
--end=-7200 \
--title="MARM01 - Po3 - vers COLT4M01" \
--rigid \
--base=1000 \
--height=120 \
--width=500 \
--alt-autoscale-max \
--lower-limit=0 \
--vertical-label="bits per second" \
--slope-mode \
--font TITLE:10: \
--font AXIS:8: \
--font LEGEND:8: \
--font UNIT:8: \
DEF:a="/var/www/cacti/rra/marm01_traffic_in_2454.rrd":traffic_in:AVERAGE \
DEF:b="/var/www/cacti/rra/marm01_traffic_in_2454.rrd":traffic_in:MAX \
DEF:c="/var/www/cacti/rra/marm01_traffic_in_2454.rrd":traffic_out:AVERAGE \
DEF:d="/var/www/cacti/rra/marm01_traffic_in_2454.rrd":traffic_out:MAX \
CDEF:cdefa=a,8,* \
CDEF:cdefd=b,8,* \
CDEF:cdefe=c,8,* \
CDEF:cdefh=d,8,* \
AREA:cdefa#EACC00FF:"Inbound"  \
GPRINT:cdefa:LAST:" Current\:%8.2lf %s"  \
GPRINT:cdefa:AVERAGE:"Average\:%8.2lf %s"  \
GPRINT:cdefd:MAX:"Maximum\:%8.2lf %s\n"  \
LINE1:cdefe#EE5019FF:"Outbound"  \
GPRINT:cdefe:LAST:"Current\:%8.2lf %s"  \
GPRINT:cdefe:AVERAGE:"Average\:%8.2lf %s"  \
GPRINT:cdefh:MAX:"Maximum\:%8.2lf %s" 

For the old server

And this:

Code: Select all

/usr/bin/rrdtool graph - \
--imgformat=PNG \
--start=-86400 \
--end=-300 \
--title='MARM01 - Po3 - vers COLT4M01' \
--rigid \
--base=1000 \
--height=120 \
--width=500 \
--alt-autoscale-max \
--lower-limit='0' \
--vertical-label='bits per second' \
--slope-mode \
--font TITLE:10: \
--font AXIS:8: \
--font LEGEND:8: \
--font UNIT:8: \
DEF:a="/appcacti/cacti/rra/marm01_traffic_in_2729.rrd":'traffic_in':AVERAGE \
DEF:b="/appcacti/cacti/rra/marm01_traffic_in_2729.rrd":'traffic_out':AVERAGE \
CDEF:cdefa='a,8,*' \
CDEF:cdefe='b,8,*' \
AREA:cdefa#EACC00FF:"Inbound"  \
GPRINT:cdefa:LAST:" Current\:%8.2lf %s"  \
GPRINT:cdefa:AVERAGE:"Average\:%8.2lf %s"  \
GPRINT:cdefa:MAX:"Maximum\:%8.2lf %s\n"  \
LINE1:cdefe#EE5019FF:"Outbound"  \
GPRINT:cdefe:LAST:"Current\:%8.2lf %s"  \
GPRINT:cdefe:AVERAGE:"Average\:%8.2lf %s"  \
GPRINT:cdefe:MAX:"Maximum\:%8.2lf %s" 

For the new one.


Actually I checked directly on the equipment and the old server has the right value!

Please Help, I have a few graphs with the same issue. Some are OK.


PS: During the upgrade I copied the database from the old server and made some changes in the new one so it could work with the new cacti version.

[FanstasticMrFox]

Hi I dont know if you fixed this, but I was seeing the same on our cacti 8.7 to 8.8 migration. What you need to change is the data-source Output Type ID to 64bit counters.

HTH

MrFox

[combellm]

Hi I dont know if you fixed this, but I was seeing the same on our cacti 8.7 to 8.8 migration. What you need to change is the data-source Output Type ID to 64bit counters.

HTH

MrFox

Thanks!
I'll check it!
I was a bit hopeless with this problem...

Should I make new graph or can I change it for all concerned data source?
I mean: Should I add "64" in the data template output type ID blank field ? In ds / traffic_in / traffic_out ?
Or just graphs that have this issue?

Thanks for your help!

[gandalf]

Hi I dont know if you fixed this, but I was seeing the same on our cacti 8.7 to 8.8 migration. What you need to change is the data-source Output Type ID to 64bit counters.

HTH

MrFox

+1
R.

[combellm]

Hi I dont know if you fixed this, but I was seeing the same on our cacti 8.7 to 8.8 migration. What you need to change is the data-source Output Type ID to 64bit counters.

HTH

MrFox

+1
R.

You said +1 so maybe you could help me with what I asked just above.
Thx

[gandalf]

I'm sorry to have skipped that question.
The root cause is, that you've used a Graph Template (in this case a specific traffic graph template) that itself refers to so called 32 bit COUNTER values in SNMP (those are the default COUNTERs that work will all snmp devices). But as your traffic is high "enough", a 32 bit COUNTER will wrap within a polling interval (thus producing "low" usage as a first glance).
To fix this, you'll have to instruct Cacti to use 64 but COUNTERs instead. Those are available e.g. with the Graph Template "In/Out Bits (64-bit Counters)", which now in turn refers to a Data Template using different OIDs being 64 bit COUNTERs (please know, that those are available with SNMP V2c only and are optional, that is, a device does not need to implement those COUNTERs).
Usually, you now will create a new traffic graph using the "In/Out Bits (64-bit Counters)" graph template and are all set.

But a forum user has made up his mind to implement the Fix64 plugin (hope I remember correctly), which should do the conversion job for you. But even now, old data will NOT be fixed in any way, so this will still look weird.

R.

[combellm]

I'm sorry to have skipped that question.
The root cause is, that you've used a Graph Template (in this case a specific traffic graph template) that itself refers to so called 32 bit COUNTER values in SNMP (those are the default COUNTERs that work will all snmp devices). But as your traffic is high "enough", a 32 bit COUNTER will wrap within a polling interval (thus producing "low" usage as a first glance).
To fix this, you'll have to instruct Cacti to use 64 but COUNTERs instead. Those are available e.g. with the Graph Template "In/Out Bits (64-bit Counters)", which now in turn refers to a Data Template using different OIDs being 64 bit COUNTERs (please know, that those are available with SNMP V2c only and are optional, that is, a device does not need to implement those COUNTERs).
Usually, you now will create a new traffic graph using the "In/Out Bits (64-bit Counters)" graph template and are all set.

But a forum user has made up his mind to implement the Fix64 plugin (hope I remember correctly), which should do the conversion job for you. But even now, old data will NOT be fixed in any way, so this will still look weird.

R.

Thanks a lot, I use now the "fix 64bit counter" plugin and it works pretty good!
I updated all my graph on Gigabit interface and it works...
Maybe my issue is linked:
http://forums.cacti.net/viewtopic.php?f=2&t=49330

Anyway thanks