Cacti 1.2.11 new install on new Centos 8.1 build not working.

[swainstm1]

I have tried and brand new install of Cacti on a brand new Centos install to troubleshoot and upgrade I am doing. And even on a pristine install and can't get even a new clean install of Cacti 1.2.11 running. No matter what path I take, everything goes smoothly until I first try to signon to the web interface to complete the install. At this point I see the login screen as expected, and can even see it has redirected to :-

http://x.x.x.x/cacti/install/install.php

But I can't login. I put in admin/admin, but it just flashes straight back to the login screen.

I have tried 2 different install guides to do this including :-

https://www.itzgeek.com/how-tos/linux/c ... hel-8.html

which installs Cacti from the repository, and

https://computingforgeeks.com/how-to-in ... -centos-8/

which installs from download from Cacti, but get the same result from both.

So I suspect there is something wrong with the Cacti 1.2.11 package.

I have another post which was Upgrade issue from version 1.1.19 to 1.2.11 (viewtopic.php?f=2&t=60901). But decided to make a new post focusing on the simplest problem which is a new install which seems to have a problem. And once this problem is solved, it might also solve my upgrade problem which I assumed was an upgrade issue. But this assumption might be wrong, and my upgrade might just work once we sort out a basic install.

I also not there is another post with issue with 1.2.11 install stopping at 42% on other platforms (ie Centos 7). I can't help if the issues are related??

Thanks everyone for your help and I look forward to a resolution.

[cigamit]

https://github.com/Cacti/cacti/issues/3436

[swainstm1]

Thanks for pointing out that thread. In trawling through that I found what appears to have fixed the problem :-

"Okay, this is interesting. Are any of you guys copying the config.php.dist to config.php? If so, comment out the $cacti_cookie_domain line. This was supposed to be commented out, but for some reason (likely late night), it was not commented out."

So I commented "$cacti_cookie_domain" line out in config.php, and then it all started working.

So for me that seems to be the fix at least for a new install.

I will try my upgrade again.

[swainstm1]

Actually, issues not completely resolved. I was able to complete the install. I was also able to signon, and confirmed the automatically added Local Linux Machine was there and polling ok. But my problem came as soon as I went to add a host. When I pressed the "Create" button, it went grey, and everything stopped there. After waiting a while and nothing happened, I cancelled out and got a warning about things not being saved. Looking in devices confirmed the new host was not there.

Looking in the cacti log file, I found :-

04/24/2020 22:15:43 - SYSTEM WARNING: Primary Admin account notifications disabled! Unable to send administrative Email.
04/24/2020 22:15:43 - POLLER: Poller[1] WARNING: The Poller Interval is 300 and you have a Data Source with a sampling interval of 60. Change your Poller Interval to 60 seconds, and repopulate your poller cache.
04/24/2020 22:15:43 - AUTOM8 NOTE: Graph Added - Device[2], Graph[5], DS[6]
04/24/2020 22:15:43 - ERROR PHP ERROR: Uncaught DivisionByZeroError: Modulo by zero in /usr/share/cacti/lib/api_poller.php:133 Stack trace: #0 /usr/share/cacti/lib/api_poller.php(87): api_poller_get_rrd_next_step('2', '60', '7') #1 /usr/share/cacti/lib/utility.php(243): api_poller_cache_item_add('2', Array, '7', '60', 2, 'uptime', 1, '/usr/share/cact...') #2 /usr/share/cacti/lib/utility.php(720): update_poller_cache(Array) #3 /usr/share/cacti/lib/api_automation.php(2261): push_out_host('2', '7') #4 /usr/share/cacti/lib/api_automation.php(2121): automation_execute_graph_template('2', '35') #5 /usr/share/cacti/lib/api_device.php(895): automation_hook_graph_template('2', '35') #6 /usr/share/cacti/lib/api_device.php(787): api_device_update_host_template('2', 5) #7 /usr/share/cacti/host.php(253): api_device_save('0', 5, 'untangle', '192.168.192.1', 'SnMp2158', '2', '', '', '161', '500', '', '2', '1', '23', '400', '1', '', 'MD5', '', 'DES', '', '', '10', '1', '1', '2', '', 'None') #8 /usr/share/cacti/host.php(60): form_save() #9 {main} thrown in file: /usr/share/cacti/lib/api_poller.php on line: 133
04/24/2020 22:15:43 - CMDPHP PHP ERROR Backtrace: (CactiShutdownHandler())

At install, I did select the option to have 1 minutes intervals. And when I check the "Poller Interval" in Settings/Poller it is set to 1 minute. Cron interval is set to 5 minutes. I assume that is all correct. But I thought I would try to clear the poller cache. But I got an error there as well. I got and error in the GUI that said "Action Failed""Sorry , we could not process you last action. Error: 500 Internal Server Error. Reason: The response to the last action was unexpected..

In the cacti error log it showed :-

04/24/2020 22:43:03 - SYSTEM WARNING: Primary Admin account notifications disabled! Unable to send administrative Email.
04/24/2020 22:43:03 - POLLER: Poller[1] WARNING: The Poller Interval is 300 and you have a Data Source with a sampling interval of 60. Change your Poller Interval to 60 seconds, and repopulate your poller cache.
04/24/2020 22:43:03 - ERROR PHP ERROR: Uncaught DivisionByZeroError: Modulo by zero in /usr/share/cacti/lib/api_poller.php:133 Stack trace: #0 /usr/share/cacti/lib/api_poller.php(87): api_poller_get_rrd_next_step('1', '60', '2') #1 /usr/share/cacti/lib/utility.php(243): api_poller_cache_item_add('1', Array, '2', '60', 1, '', 1, 'perl /usr/share...') #2 /usr/share/cacti/lib/utility.php(67): update_poller_cache(Array) #3 /usr/share/cacti/utilities.php(40): repopulate_poller_cache() #4 {main} thrown in file: /usr/share/cacti/lib/api_poller.php on line: 133
04/24/2020 22:43:03 - CMDPHP PHP ERROR Backtrace: (CactiShutdownHandler())

This is a very default installation as previous outlined on a brand new box.

Any ideas how to fix?

[swainstm1]

OK, more fiddling with things, and I think I might have worked around the previous issue, but still not completely there. I tried changing the poller interval to 5 minutes. That still did not work with all the same message. So I changed in back to 60 seconds. I think tried clearing the poller cache, which then worked. So then tried adding a device, and it all worked.

So for me that is the workaround. Change the poller cache from 60 sec to 5 minutes and then back again. But I assume this highlights a little bug that might be work fixing that has something to do with the polling cache value appearing to be set correctly, but in practice it is not. And it appears that changing to there and back fixes the issue??

[swainstm1]

So now I appear to be able to create a host and pressing create gets me to a successful message. However, I can't see the host if I go management / devices. I can see the local linux machine that was created at install, but not the hosts that were added??? Would I be right in assuming they should show up there or is there some issue with the way I have created the host. The host created are in the system somehow, but not visible through Managerment/Devices by default. But I can see the number of hosts have increased in the poller cache, I can see the devices when look at the tree view and I can see available devices.

So fiddling around, I found that I could find the devices if I went to management and devices and did a search for the device name, then it found the device. But for some reason, they do not show up without this?? But if I clear the search history, it disappears from the list. I have a bunch of devices that were automatically discovered all in the list. But it seems the devices I create do not show up?

I seem to be having a lot of little issues, that there does not seem to be easy intuitive ways around. With a lot of fiddling and trial and error I have been able to workaround some of these issue. But it does not seem right, and I suspect there is still something wrong with my install that is creating all these little gremlins?? I am bit surprised by this, as I have been using Cacti for a long time (earlier versions), and while there are sometimes issues to workaround, generally I have found the stability to be quite good. But 1.2.11 seems to be very difficult.

Any ideas?

[TheWitness]

That's actually a pretty good finding. Might have to log this issue. Will trace the error message.