Cacti 0.8.6d Out

[Wernicke]

sure, but it's working fine now with that PATH string added in /etc/crontab

Still don't understand why cacti forgets this.. it does not matter if i use CMD or CACTID

-Sigurd

[fletch]

177 of graphs are updating fine after upgrading.
The other 700+ are not updating.

I have DEBUG logging on and the rrdtool update commands in the log are not logging any errors.

I copy and paste one of the rrdtool update commands and run it manually and the rrd file updates.

I am looking into what is different about these graphs that is causing rrdtool update to not update and to fail silently.

I upgraded from rrdtool1.0.x to 1.2.1 - same graphs not updating.

I tried adding to the cron PATH, and wrapping in a script for cron - same prob.


I'm going to start digging into the graph templates next....arg

[fletch]

I'm going to revert to the previous version - but I think it may be a templates issue - I see this difference in the logs:

Old working server log:

RRDCMD: update 'rra/irtweb04_apache_idle_workers_1151.rrd' --template apache_
total_hits:apache_total_kbytes:apache_busy_workers:apache_idle_workers N195525:17:23

(notice the N: in the output)

New - non working example:

05/02/2005 10:39:51 AM - POLLER: Poller[0] CACTI2RRD: /usr/local/bin/rrdtool update rra/irtweb04_apache_idle_workers_1151.rrd --template apache_total_hits:apache_total_kbytes:apache_busy_workers:apache_idle_workers 1115055575:578504:7726968:3:37


The apache script is returning the same stuff - but cacti is passing it differently to rrdtool...

dunno why at this point...

[TheWitness]

Did you install from RPM. A mistake that many make is having 2 scheduled crons running at the same time. This screws up the polling.

The value in the rrdtool update is not an issue. There is a revised lib/poller.php that I can send you that may have some affect. Please let me know if that is what you want to do.

TheWitness

[fletch]

I did not install from RPM -
Sure I'll try the updated poller

Thanks.

[TheWitness]

Not posted yet. But you can edit and locate the following line:

"order by poller_output.time"

Please remove the order by clause while maintaining appropriate syntax and you will be patched.

TheWitness

[fletch]

Removed the order by :
/* create/update the rrd files */
$results = db_fetch_assoc("select
poller_output.output,
poller_output.time,
poller_output.local_data_id,
poller_item.rrd_path,
poller_item.rrd_name,
poller_item.rrd_num
from poller_output,poller_item
where (poller_output.local_data_id=poller_item.local_data_id and poller_output.rrd_name=poller_item.rrd_name)");

Now, I am getting hanging pollers:

05/02/2005 05:12:55 PM - POLLER: Poller[0] DEBUG: About to Spawn a Remote Process [CMD: , ARGS: 0 64]
Waiting on 1/1 pollers.
Waiting on 1/1 pollers.
Waiting on 1/1 pollers.
Waiting on 1/1 pollers.
Waiting on 1/1 pollers.
Waiting on 1/1 pollers.
Waiting on 1/1 pollers.
Waiting on 1/1 pollers.
Waiting on 1/1 pollers.
Waiting on 1/1 pollers.
Waiting on 1/1 pollers.
Waiting on 1/1 pollers.

[TheWitness]

Can you please run cmd.php from the command line in debug mode. And then please post output.

TheWitness

[fletch]

output from php cmd.php (attached)

[fletch]

I disabled the device that was often the last rrd to update and now all the rrds are updating !!!
Yayaya

But this is with cmd.php and it takes 10 mins

cactid 0.8.6e is core dumping

I'm going to try d again...

[fletch]

nope - back to cmd.php

cactid D is only finishing 137 rrds and now its hanging .

Will have to live with cmd.php every 10 mins...

[fletch]

BTW, I notice the script server is running when it hangs:

cacti 10436 1 0 14:53:44 pts/5 0:00 /usr/local/cactid/bin/cactid 0 64
cacti 10439 10436 0 14:53:44 pts/5 0:00 /usr/local/bin/php /fs/web/irt-web/htdocs/irt-cacti2/script_server.php cactid 0

[TheWitness]

You have a messed up database. Please edit Data Sources and select Host = None and delete the bogus entries and you should be back in business with Cactid. I will look at another change in the mean time (Cactid 0.8.6e RC3) to adjust for this issue.

TheWitness

[TheWitness]

Fletch,

Please state:

Number of Hosts
Number of RRD Files
Script/Snmp Mix

TheWitness

[fletch]

I deleted the 3 entries listed with template "none"

trying cactid again...