Ad blocker detected: Our website is made possible by displaying online advertisements to our visitors. Please consider supporting us by disabling your ad blocker on our website.
Marco wrote:But cactid seems not to know where the configuration is.
By default, cactid searches the local directory for cactid.conf. It will print out where it found cactid.conf (at least when running in DEBUG mode). There are some fixed locations which are searched as well, amongst them is /etc/cactid.conf which I do prefer.
Reinhard
I just noticed I was having a similar problem with my load graphs. I reverted to using cactid-0.8.6e and things are working again. Was there perhaps new code/changed code that was introduced in 0.8.6f that is creating this problem? I maintain the cactid package in Gentoo, and I'm half tempted to p.mask this version (I'm surprised I haven't seen this in our bugzilla yet).
Let me know when you figure out what the problem is.
lvm wrote:
By default, cactid searches the local directory for cactid.conf. It will print out where it found cactid.conf (at least when running in DEBUG mode). There are some fixed locations which are searched as well, amongst them is /etc/cactid.conf which I do prefer.
Reinhard
I know that, but I get following error:
www-data@FRY:~$ /usr/sbin/cactid
CACTID: MYSQL: Connection Failed: Access denied for user: 'cactiuser@localhost' (Using password: YES)
If I change the directory to /etc/cacti then there are no Problems.
chadd wrote:I was running Cactid with 5 threads, but I changed it to 1 thread and get the same results.. Hosts go down, and graphs do not update.
I am willing to try, with some further explaination, marco's suggestion. I am a little unclear as to what he wants me to do. Thanks for any help.
I had no such Problems with cactid and segfaults or something else.
The Problem was that cactid doesn't take his config from /etc/cacti/.
And yes with 1 Thread its also running.
There is a known problem with Cactid 0.8.6f and some OS' with SNMP results that are handled as a string. The problem is also causing "some" people problem with script results.
I am trying to identify the section of code that works in some OS' and not others. When I have that information, and resolve it, I plan on releasing Cactid 0.8.6g.
TheWitness
True understanding begins only when we realize how little we truly understand...
In the mean time, I continue to seek volunteers to help me resolve this issue. If you are having problems with either and can give me RO access to your agents, it would be greatly appreciated.
TheWitness
True understanding begins only when we realize how little we truly understand...
Once I switched my "Downed Host Detection" method to "SNMP -Reliable" from "PING and SNMP - Most Recommended" everything worked for me again and I was able to crank my thread level up on cactid.
drose12 wrote:Once I switched my "Downed Host Detection" method to "SNMP -Reliable" from "PING and SNMP - Most Recommended" everything worked for me again and I was able to crank my thread level up on cactid.
I tried that, and it worked. Cactid is now working, and updating all my graphs.
If I change it back to snmp and ping, then only a few of the devices go down, but they do go down and do not graph. ...Change it back(snmp only), they recover again, and start to graph again.
Does this give any clues as to why Cactid demonstrates this behavior?
I could get Cactid to poll from the command line.
1. Set the option to use cactid in the web interface
2. Run the command by hand from a prompt: /usr/local/bin/php /opt/apache/htdocs/cacti/poller.php > /dev/null 2>&1
Note: See I put the full path to PHP just like in my crontab:
*/5 * * * * root /usr/local/bin/php /opt/apache/htdocs/cacti/poller.php > /dev/null 2>&1
It ran great from the command line.
I would walk away and NO GRAPHS 10 minutes later.
I realized running as root under cron, cactid could not find where my libraries where located.
I added this to the top of /etc/crontab to set environment variables: LD_LIBRARY_PATH=/usr/local/lib:/opt/net-snmp/lib/:/opt/mysql/lib/mysql/
Ah, now cactid knows how to find net-snmp and mysql libs.
Graphing fixed.
N3NCY wrote:Just a thought, here is what worked for me:
I could get Cactid to poll from the command line.
1. Set the option to use cactid in the web interface
2. Run the command by hand from a prompt: /usr/local/bin/php /opt/apache/htdocs/cacti/poller.php > /dev/null 2>&1
Note: See I put the full path to PHP just like in my crontab:
*/5 * * * * root /usr/local/bin/php /opt/apache/htdocs/cacti/poller.php > /dev/null 2>&1
It ran great from the command line.
I would walk away and NO GRAPHS 10 minutes later.
I realized running as root under cron, cactid could not find where my libraries where located.
I added this to the top of /etc/crontab to set environment variables: LD_LIBRARY_PATH=/usr/local/lib:/opt/net-snmp/lib/:/opt/mysql/lib/mysql/
Ah, now cactid knows how to find net-snmp and mysql libs.
Graphing fixed.
Hopes this helps somebody!
I tried your suggestion, and it did not work. I had the same results - many hosts go to a 'down' state, and do not graph.
chadd wrote:If I change it back to snmp and ping, then only a few of the devices go down, but they do go down and do not graph. ...Change it back(snmp only), they recover again, and start to graph again.
Does this give any clues as to why Cactid demonstrates this behavior?
-chadd
As to SNMP checking, well, I think this is clear (UDP:161 check). "Ping" will default to UDP Pinging some port above 30,000 (tcpdump will help figuring this out). If the remote host does not respond to this (e.g. some firewall doing its job...), cactid will assume the host is dead. And therefor he will not ask him anymore, so: no graphs.
Options: make UDP Ping work with your remote host, OR simply use "SNMP - Reliable".
Reinhard
Ditto here.. Having the same issue. I have 0.8.6g, all the pre-patches installed, and cactid 0.8.6f-1. Everything is compiled against net-snmp 5.2.1.2.
Cmd.php works fine. albeit it takes a while and causes some major load.
Cactid looks like it's running but doesn't generate any graphs. Running it in debug mode only generate a few errors "WARNING: Result from SNMP not valid. Partial Result: ..."
While those same checks using cmd.php work fine gathering the SNMP data.
tmoore wrote:Ditto here.. Having the same issue. I have 0.8.6g, all the pre-patches installed, and cactid 0.8.6f-1. Everything is compiled against net-snmp 5.2.1.2.
Cmd.php works fine. albeit it takes a while and causes some major load.
Cactid looks like it's running but doesn't generate any graphs. Running it in debug mode only generate a few errors "WARNING: Result from SNMP not valid. Partial Result: ..."
While those same checks using cmd.php work fine gathering the SNMP data.
The load issue, I think I can help with... If you don't have php-snmp installed, do so, that should take care of at least some of the load problem. Other than that, I am still waiting on a fix. I am however using cactid now, just specifying snmp only instead of ping and snmp.. That seems to work. Also, make sure you are using cactid-f-1, not f.
