| Author |
Message |
|
EoinDubh
User
Joined: Thu Aug 16, 2007 11:24 am Posts: 51
|
 Cacti seems to stop gathering data at tines
My Cacti 0.3 system seems to just stop gathering data for from 15 minutes to several hours at different times. I have attached a graphic of a chart showing a couple of gaps. The one at the left was 4 hours long. I do not see any settings in crontab that would account for this. This happens with all graphs for all devices. Anyone have any Ideas?
Attachments:
File comment: Graphic attachment
graph_image.php.png [22.51 KiB]
Downloaded 1650 times
File comment: localhost chart
localhost_image.php.png [ 31.63 KiB | Viewed 980 times ]
|
| Thu Nov 29, 2007 9:50 am |
|
 |
|
jimmy
Site Admin
Joined: Thu Nov 11, 2004 11:12 am Posts: 2007
|
Anything in the Cacti log?
|
| Sat Dec 01, 2007 11:06 pm |
|
 |
|
EoinDubh
User
Joined: Thu Aug 16, 2007 11:24 am Posts: 51
|
Nothing in the log. I looked through all the logs that I could find on the machine and do not see anything.
|
| Sun Dec 02, 2007 5:24 pm |
|
 |
|
lonecrow
User
Joined: Tue Jan 15, 2008 10:45 am Posts: 27
|
I get this really bad for all my graphs on a default install of cactiez.
I wish I knew how to update to 0.8.7a from this version...
Can't find any good instructions anywhere.
|
| Mon Jan 28, 2008 10:28 am |
|
 |
|
mrloon
New User
Joined: Fri Jan 25, 2008 1:58 pm Posts: 4
|
You should check that your logging is set to show errors/warnings during the poll. Least then you should be able to see any issues when it is running the poll. I'm assuming that you are using cmd.php not spine when running the default poll correct? If not, set it to cmd for now, set logging to Debug level, run a poll manually, set logging back to your prior setting and view the log. If there is issues getting the initial data, you should see errors/warnings there. That should at leave give a starting point where to look for issues.
|
| Mon Jan 28, 2008 2:24 pm |
|
 |
|
Linuxx
New User
Joined: Tue Jan 29, 2008 1:32 pm Posts: 16
|
Edit your crontab
#crontab -e
change
*/1 * * * * php /var/www/html/poller.php > /dev/null 2>&1
To say
*/5 * * * * php /var/www/html/poller.php > /dev/null 2>&1
This will fix your issue.
Linuxx
|
| Tue Jan 29, 2008 3:31 pm |
|
 |
|
EoinDubh
User
Joined: Thu Aug 16, 2007 11:24 am Posts: 51
|
In my case I already have */5 * * * * root /usr/bin/php /opt/cacti/poller.php > /dev/null 2>&1 in /etc/crontab and it made no difference. I think that something was getting corrupted over time as graphing stopped completely at midnight last Thursday. I am bringing down the VM and creating another from the ISO.
|
| Tue Jan 29, 2008 3:57 pm |
|
 |
|
Linuxx
New User
Joined: Tue Jan 29, 2008 1:32 pm Posts: 16
|
I am running in VMWARE too. I would give you my VMDK if it didn’t have so much proprietary info in it! :D
I had the same problem as you, and had to monkey with the poller settings. I have 1600 total data sources. Here are my settings,
Good luck!
Linuxx
Attachments:
poller.JPG [ 123.56 KiB | Viewed 1035 times ]
|
| Tue Jan 29, 2008 4:05 pm |
|
 |
|
jimmy
Site Admin
Joined: Thu Nov 11, 2004 11:12 am Posts: 2007
|
You have to update to 0.8.7a to fix this, since that was a bug in 0.8.7 (multiple pollers launching) which causes this issue. Just follow the normal upgrade path for Cacti, then install the PA like its a new install.
|
| Mon Feb 04, 2008 1:20 am |
|
|