Post 0.8.6.f migration - Cannot create graphs

[havoc1942]

I have recently migrated from 0.8.6.d to 0.8.6.f and cactid 0.8.6.d to 0.8.6.e and find myself with a couple of problems.

1. I cannot create new graphs for any device, but I can create new devices.
2. All of my Netscreen firewalls are not being polled. The devices show down, but they respond to SNMP when I edit the devices.

FC2
php -m has mysql support
rrd 1.0.49
net-snmp-5.1.1-2
cacti 0.8.6.f
cactid 0.8.6.e

Thanks!

[TheWitness]

What is your availability method set to.

TheWitness

[havoc1942]

Ping and SNMP with UDP Ping "Ping Type".

Just to detail....
When I create a device and move to the "Create Graphs" screen I can select all the items on the device available for graphing, but when I hit submit it just returns me to the same screen with nothing selected and doesn't create and .rrd files.

Thanks.

[TheWitness]

Check your Apache error log for clues. Also, do you have built-in snmp support in PHP? What version of PHP are you running?

TheWitness

[havoc1942]

PHP is not using built in SNMP support (extension=...)

PHP version 4.3.10

Here is the content of the error_log file from httpd. Every time I try to add a new graph I get the errors.

PHP Warning: preg_match: internal pcre_fullinfo() error -3 in /var/www/html/cacti/graphs_new.php on line 70, referer: http://10.96.10.32/graphs_new.php?host_id=452
PHP Warning: preg_match: internal pcre_fullinfo() error -3 in /var/www/html/cacti/graphs_new.php on line 64, referer: http://10.96.10.32/graphs_new.php?host_id=452
PHP Warning: preg_match: internal pcre_fullinfo() error -3 in /var/www/html/cacti/graphs_new.php on line 66, referer: http://10.96.10.32/graphs_new.php?host_id=452
PHP Warning: preg_match: internal pcre_fullinfo() error -3 in /var/www/html/cacti/graphs_new.php on line 70, referer: http://10.96.10.32/graphs_new.php?host_id=452

Thanks.

[raX]

Searching for this error on Google reveals that many people are experiencing this issue. The problem appears to be within PHP itself, as there is already a bug filed. It looks like it has to due with using an external PCRE library versus the one that is bundled with PHP. If possible, I would try recompiling/upgrading PHP to see if that makes a difference.

-Ian

[havoc1942]

Fixed...

I had to rebuild php with the following in the configure options ( --with-pcre-regex) to use the php pcre libraries.

Thanks

[havoc1942]

I thought everything was good, but I still have four devices that have issues. In the device management they show down, but have 90%+ availability and show current response time within range. If I select the device it shows SNMP data. Yet, when I look at the graphs they are all nan. I have found no entries in the cacti debugged log for these hosts. I have verified that the devices respond to SNMP from the command line.

These are the only devices that are having this problem.

Thanks.

[raX]

It seems to me that your devices are not being polled because the poller thinks they are down for some reason. Turning up the poller logging level might idicate why. There are two things that I would try.

1) Clear your poller cache under System Utilities on the menu.
2) Change your Downed Host Detection setting to "SNMP - Reliable".

-Ian