Hey look a useless thread where I ramble about my problems

[sirchode]

I'm running Windows 2000 with MySQL 5.0.17 and PHP 5.1.1. I followed the updated Windows installation instructions, including the creation of the Scheduled Task for polling.php. Every five minutes (and when I manually invoke the task), I am greeted with this dialog box and the typical Windows "ding":



It's incredibly annoying. I've looked through polling.php trying to see if there's anything I could modify manually but didn't find anything. Any idea what would cause this and what the solution might be?

[gandalf]

MySQL 5.0 is not supported until cacti-0.8.6h which is currently beta
Reinhard

[sirchode]

MySQL 5.0 is not supported until cacti-0.8.6h which is currently beta
Reinhard

I figured this would be part of the issue, so I went ahead and rolled back to MySQL 4.1.16. The problem still persists. Any ideas as to what's causing it?

[sirchode]

Now that my other problem regarding graph creation has been solved, I've found that I can't view any graphs I create. I can see that there are no .rra files from which graphs are made, so that leads me to believe this poller problem is somehow related to the lack of .rra files.

While in cacti, I can click on a device and see its SNMP information, so my only guess is that it has something to do with poller.php and this scheduled task not being able to retrieve the data.

[gandalf]

Did you follow the instructions of BSOD2000 at http://bsod2600.home.comcast.net/install_windows.html?
Reinhard

[sirchode]

Did you follow the instructions of BSOD2000 at http://bsod2600.home.comcast.net/install_windows.html?
Reinhard

As a matter of fact, I followed the "old" instructions that were linked on the main download page (that should really be updated), but when I noticed his newer guide, I went through it step-by-step again and made any changes that were different. As far as I know, I've done everything in the updated guide.

I've been tinkering with this non-stop and I've managed to get to the point where the error doesn't pop up (and I forget what I did to make it go away), but Task Scheduler now says "Could not start" about the poller.php task. I'm looking at my PHP configuration and comparing it to BSOD2600's updated guide and seeing if I missed anything.

I added C:\php to the PATH variable but for some reason I still couldn't go to a command prompt, navigate to the cacti directory and type "php poller.php". So I navigated to C:\php and typed "php C:\<Apache path>\cacti\poller.php", and it appeared to work. I eventually saw this:

12/29/2005 12:15:53 PM - POLLER: Poller[0] Maximum runtime of 296 seconds exceeded. Exiting.

So yeah, I guess I'm making some sort of progress, but I still feel like a blind man walking

[sirchode]

Alright, I'm so close to this I can taste it. Here are the results of running poller.php from the command line:

--------------------------------------------
C:\Program Files\Apache Group\Apache2\cacti>php poller.php
ERROR: can't parse argument 'Files/Apache'
ERROR: opening 'C:\Program': No such file or directory
ERROR: can't parse argument 'Files/Apache'
ERROR: opening 'C:\Program': No such file or directory
ERROR: can't parse argument 'Files/Apache'
ERROR: opening 'C:\Program': No such file or directory
X-Powered-By: PHP/4.4.1
Content-type: text/html

12/29/2005 04:33:04 PM - SYSTEM STATS: Time:2.3976 Method:cactid Processes:1 Threads:1 Hosts:2 HostsPerProcess:2 DataSources:5 RRDsProcessed:3
--------------------------------------------

And here is what shows up in cacti.log:

--------------------------------------------
12/29/2005 04:33:03 PM - CACTID: Poller[0] Host[2] SNMP Result: Host responded to SNMP
12/29/2005 04:33:03 PM - CACTID: Poller[0] Host[2] SNMP: v1: 192.168.0.3, dsname: 5min_cpu, oid: .1.3.6.1.4.1.9.2.1.58.0, value: 0
12/29/2005 04:33:03 PM - CACTID: Poller[0] Host[2] SNMP: v1: 192.168.0.3, dsname: traffic_in, oid: .1.3.6.1.2.1.2.2.1.10.2, value: 125211
12/29/2005 04:33:03 PM - CACTID: Poller[0] Host[2] SNMP: v1: 192.168.0.3, dsname: traffic_out, oid: .1.3.6.1.2.1.2.2.1.16.2, value: 131282
12/29/2005 04:33:03 PM - CACTID: Poller[0] Host[2] SNMP: v1: 192.168.0.3, dsname: traffic_in, oid: .1.3.6.1.2.1.2.2.1.10.1, value: 110628212
12/29/2005 04:33:03 PM - CACTID: Poller[0] Host[2] SNMP: v1: 192.168.0.3, dsname: traffic_out, oid: .1.3.6.1.2.1.2.2.1.16.1, value: 16579050
12/29/2005 04:33:03 PM - CACTID: Poller[0] Time: 0.8020 s, Threads: 1, Hosts: 1
12/29/2005 04:33:04 PM - SYSTEM STATS: Time:2.3976 Method:cactid Processes:1 Threads:1 Hosts:2 HostsPerProcess:2 DataSources:5 RRDsProcessed:3
--------------------------------------------

Unfortunately, this means there are still no .rrd files in cacti/rra, which means still no graphs. I'm going to take a guess that cacti is the culprit here, I thought I remember reading it doesn't yet support long filenames. The problem is that I've had Apache installed on here forever, so I might have to uninstall it and reinstall to C:\apache.

[sirchode]

Sorry for the blog-like thread but now I'm down to just this error after reinstalling Apache:

---------------------------------
C:\apache\Apache2\cacti>php poller.php
ERROR: opening 'C:\Program': No such file or directory
ERROR: opening 'C:\Program': No such file or directory
ERROR: opening 'C:\Program': No such file or directory
X-Powered-By: PHP/4.4.1
Content-type: text/html

12/29/2005 05:01:58 PM - SYSTEM STATS: Time:1.0372 Method:cactid Processes:1 Thr
eads:1 Hosts:2 HostsPerProcess:2 DataSources:5 RRDsProcessed:3
---------------------------------

I have gone through the guides and changed everything I could think of back to c:\apache\Apache2\ but it looks like I'm still missing something. Anyone know where I need to edit?

I also still can't get the Scheduled Task to actually work. I'm not too worried about that right now, though, we'll see if it's still being stubborn once this probem is solved.

[sirchode]

Well all of my errors are fixed and resolved except one new one. I have no idea how or why the errors in my previous post stopped, but they eventually did. As for the Task Scheduler not being able to start, I had been trying to run the task as SYSTEM (as mentioned in the sticky thread). I decided to give it a shot as a local Administrator and sure enough it worked like a charm, so now things are great.

This newest problem is kinda strange and I doubt I'll receive a reply about it at this point. When I click on Devices, then click on a device, at the top of the page I see "SNMP Error" in red:



The strange thing is that this wasn't always the case, just earlier today it showed all the proper SNMP information. I have no idea what changed, but I can't get it back. This also means I can't create interface graphs on newly-added devices, so it's pretty crappy. I logged into one of the routers and debugged the SNMP packets and it looks like this:

Code: Select all

littlefoot#debug snmp packets
SNMP packet debugging is on
littlefoot#
Dec 29 23:30:02: SNMP: Packet received via UDP from 192.168.0.101 on Ethernet0
Dec 29 23:30:02: SNMP: Get request, reqid 536007921, errstat 0, erridx 0
 system.1.0 = NULL TYPE/VALUE
Dec 29 23:30:02: SNMP: Response, reqid 536007921, errstat 0, erridx 0
 system.1.0 = Cisco Internetwork Operating System Software
Dec 29 23:30:02: SNMP: Packet sent via UDP to 192.168.0.101
Dec 29 23:30:02: SNMP: Packet received via UDP from 192.168.0.101 on Ethernet0
Dec 29 23:30:02: SNMP: Get request, reqid 536007922, errstat 0, erridx 0
 sysUpTime.0 = NULL TYPE/VALUE
Dec 29 23:30:02: SNMP: Response, reqid 536007922, errstat 0, erridx 0
 sysUpTime.0 = 61874519
Dec 29 23:30:02: SNMP: Packet sent via UDP to 192.168.0.101
Dec 29 23:30:02: SNMP: Packet received via UDP from 192.168.0.101 on Ethernet0
Dec 29 23:30:02: SNMP: Get request, reqid 536007923, errstat 0, erridx 0
 lsystem.58.0 = NULL TYPE/VALUE
Dec 29 23:30:02: SNMP: Response, reqid 536007923, errstat 0, erridx 0
 lsystem.58.0 = 0
Dec 29 23:30:02: SNMP: Packet sent via UDP to 192.168.0.101
Dec 29 23:30:02: SNMP: Packet received via UDP from 192.168.0.101 on Ethernet0
Dec 29 23:30:02: SNMP: Get request, reqid 536007924, errstat 0, erridx 0
 ifEntry.10.2 = NULL TYPE/VALUE
Dec 29 23:30:02: SNMP: Response, reqid 536007924, errstat 0, erridx 0
 ifEntry.10.2 = 1273475
Dec 29 23:30:02: SNMP: Packet sent via UDP to 192.168.0.101
Dec 29 23:30:02: SNMP: Packet received via UDP from 192.168.0.101 on Ethernet0
Dec 29 23:30:02: SNMP: Get request, reqid 536007925, errstat 0, erridx 0
 ifEntry.16.2 = NULL TYPE/VALUE
Dec 29 23:30:02: SNMP: Response, reqid 536007925, errstat 0, erridx 0
 ifEntry.16.2 = 1311405
Dec 29 23:30:03: SNMP: Packet sent via UDP to 192.168.0.101
Dec 29 23:30:03: SNMP: Packet received via UDP from 192.168.0.101 on Ethernet0
Dec 29 23:30:03: SNMP: Get request, reqid 536007926, errstat 0, erridx 0
 ifEntry.10.1 = NULL TYPE/VALUE
Dec 29 23:30:03: SNMP: Response, reqid 536007926, errstat 0, erridx 0
 ifEntry.10.1 = 115354857
littlefoot#
Dec 29 23:30:03: SNMP: Packet sent via UDP to 192.168.0.101
Dec 29 23:30:03: SNMP: Packet received via UDP from 192.168.0.101 on Ethernet0
Dec 29 23:30:03: SNMP: Get request, reqid 536007927, errstat 0, erridx 0
 ifEntry.16.1 = NULL TYPE/VALUE
Dec 29 23:30:03: SNMP: Response, reqid 536007927, errstat 0, erridx 0
 ifEntry.16.1 = 17650146
Dec 29 23:30:03: SNMP: Packet sent via UDP to 192.168.0.101
littlefoot#no debug all
All possible debugging has been turned off
littlefoot#

And here is what cacti shows in its debug section:



The graphs are updated regularly and accurately via these SNMP exchanges, but for some reason cacti just can't seem to get what it needs for the Device list. The NULL TYPE/VALUE parts look a little strange, but I don't know the SNMP protocol well enough to make sense of it.

I have enabled SNMP traps and informs and searched the forums looking for answers. Someone had a similar issue earlier, his suggestion was to increase the timeout limit but this hasn't worked so far for me.

So, one last time, does anyone have any ideas?

[TheWitness]

1) of course you know now that paths with spaces are not supported.
2) edit the settings table and look for abarrent paths:

Code: Select all

mysql -u root -p cacti
>select * from settings where name like '%path%';

3) if you find some ugly paths, change them using something like MySQL Query Browser
4) clear your poller cache

TheWitness

[sirchode]

Code: Select all

mysql> select * from settings where name like '%path%';
+---------------------------+---------------------------------------+
| name                      | value                                 |
+---------------------------+---------------------------------------+
| path_rrdtool              | c:/rrdtool/rrdtool.exe                |
| path_php_binary           | c:/php/php.exe                        |
| path_cactilog             | C:/apache/Apache2/cacti/log/cacti.log |
| path_snmpwalk             |                                       |
| path_snmpget              |                                       |
| path_rrdtool_default_font | C:/WINNT/Fonts/Vera.ttf               |
| path_cactid               | c:/cactid/cactid.exe                  |
| path_webroot              | c:/apache/Apache2/cacti               |
| path_html_export          |                                       |
| path_html_export_skip     |                                       |
+---------------------------+---------------------------------------+
10 rows in set (0.02 sec)

All I can say is that the paths somehow corrected themselves. I was literally hitting Up+Enter and taking screengrabs to try to read some of the error messages when I noticed it had changed and started working, I don't know what to say about that.

Clearing my poller cache doesn't seem to work. As in, when I click Clear Poller Cache it reloads and shows me a list of 6 items in the cache. I can click and click but it doesn't seem to actually clear out. Clearing the Cacti Log File works, though.

[TheWitness]

"Clearing" is a little misleading. It should more appropriately state "Rebuilding". The poller grabs and sets the path_webroot automatically. Clearing the poller cache will correct some script paths related to that incorrect path.

TheWitness