LOST -Solved

[rtyrell]

Upgraded from cacti-0.8.5a to cacti-0.8.6g fully patched
*cactid-0.8.6f-1.
*Red Hat Enterprise Linux WS release 3 (Taroon Update 5)
*PHP 4.3.2 (cgi), Zend Engine v1.3.0,
*RRDtool 1.0.50
*mysql-3.23.58-15.RHEL3.1
*php memory limit set to 256MB
*truncated poller_output/ cleared poller cache
* crontab= */5 * * * * noc /opt/cacti/poller.php
noc=cactiuser

I'm monitoring approximately 70 hosts. For some reason about 20 hosts will not update their rrd files. DEBUG shows the rrd files getting updated but they never do. After researching numerous posts I have tried just about everything. Permissions are the same for all rrd files. I am also unable to graph/rrd new hosts. I'm LOST! Any help would be appreciated.

I have ran cactid manually and it shows the rrdtool update, yet no update happens (for specific hosts/ new hosts since upgrade).


Thanks!

[gandalf]

Upgraded from cacti-0.8.5a to cacti-0.8.6g fully patched
*cactid-0.8.6f-1.
*Red Hat Enterprise Linux WS release 3 (Taroon Update 5)
*PHP 4.3.2 (cgi), Zend Engine v1.3.0,
*RRDtool 1.0.50
*mysql-3.23.58-15.RHEL3.1
*php memory limit set to 256MB
*truncated poller_output/ cleared poller cache
* crontab= */5 * * * * noc /opt/cacti/poller.php
noc=cactiuser

I'm monitoring approximately 70 hosts. For some reason about 20 hosts will not update their rrd files. DEBUG shows the rrd files getting updated but they never do. After researching numerous posts I have tried just about everything. Permissions are the same for all rrd files. I am also unable to graph/rrd new hosts. I'm LOST! Any help would be appreciated.

I have ran cactid manually and it shows the rrdtool update, yet no update happens (for specific hosts/ new hosts since upgrade).


Thanks!

No final solution, but...
change crontab to noc /opt/cacti/poller.php > /opt/cacti/log/poller.log 2>&1 to always get the full output of the last poller run. cacti.log would not show failing rrdtool updates (it does show only the command it _tries_ to run). To debug a single host, run cactid against that very host (you will have to knwo the host id, e.g. go to the device screen for that host and look at your URL to get the id): cactid 12 12 for host 12.
Your attachment includes the names of the failing hosts, but not their ids. So this is not useful (it would be nice to attach as a pure text file, no .doc).
Reinhard

[rtyrell]

The crontab has been changed per your suggestion. I have also re-uploaded the logfile as .txt and listed the host ids for the hosts not receiving rrd updates (via select * from host;)

I ran cactid against a known host that updates it's rrds (host 43) and against a host that is not updating it's rrds (host 107).

[root@ns-stat cactid]# ./cactid 43 43
CACTID: Using cactid config file [cactid.conf]
CACTID: Version 0.8.6f starting
CACTID: Execution Time: 15.2571 s, Threads: 5, Hosts: 2

[root@ns-stat cactid]# ./cactid 107 107
CACTID: Using cactid config file [cactid.conf]
CACTID: Version 0.8.6f starting
CACTID: Execution Time: 0.8445 s, Threads: 5, Hosts: 2

I did not receive any output, nor did the rrd files get updated for either host. That doesn't seem right.

I attached the poller.log file, which does have some error message at the end "read_config_store open failure on /var/net-snmp/cactid.conf".


Thanks for looking into this.

[gandalf]

Where did you place your cactid.conf? I'd like to have it in /etc/cactid.conf, not in the dir of cactid itself
Reinhard

[rtyrell]

I don't mind moving cactid.conf to /etc --- how does cactid know how to find it?

[gandalf]

Well, AFAIK it searches in some pre-defined (aka: coded) places. To be sure, you will have to look at the code. I'm sure it searches current dir and /etc. But there are more places ...
Reinhard

[rtyrell]

I went ahead and performed the upgrade again...and everything worked fine? Then, after installing Advanced Ping Template 1.3 for Cacti 0.8.6g with the sudo option, the problem came back. If I applied the template to atb-ds20 it caused the hosts above it (numerical value) to not update their rrds. I'm still investating the issue. Must be some bad syntax with the sudo workaround I'm using trying to get Advanced Ping to work. Thanks again.

-Rick