SNMP Time was not numeric
Moderators: Developers, Moderators
SNMP Time was not numeric
I recently upgraded cacti to 0.8.7b
Since then, I've been seeing messages such as: "Device does not require SNMP", "SNMP Time was not numeric", and "Host did not respond to SNMP".
The following is an extract from the log. Advice please. I'm not sure what they relate, and they seem contradictory.
Thanks.
02/16/2008 10:55:02 PM - CMDPHP: Poller[0] DEBUG: SQL Exec: "update host set status = '3', status_event_count = '0', status_fail_date = '2008-02-16 21:45:03', status_rec_date = '2008-02-16 22:45:03', status_last_error = 'Host did not respond to SNMP', min_time = '0.00000', max_time = '10.67000', cur_time = '0', avg_time = '0.0004199967864357', total_polls = '131634', failed_polls = '938', availability = '99.287418144248' where hostname = '127.0.0.1'"
02/16/2008 10:55:02 PM - CMDPHP: Poller[0] Host[1] NOTICE: HOST EVENT: Host Returned from DOWN State:
02/16/2008 10:55:02 PM - CMDPHP: Poller[0] Host[1] SNMP: Device does not require SNMP
02/16/2008 11:00:04 PM - POLLER: Poller[0] WARNING: Poller[0] Host[1] SNMP Time was not numeric
02/16/2008 10:55:02 PM - CMDPHP: Poller[0] DEBUG: SQL Exec: "UPDATE poller_item SET rrd_next_step = rrd_step - 300 WHERE (rrd_next_step < 0 AND host_id >= 0 AND host_id <= 3)"
Since then, I've been seeing messages such as: "Device does not require SNMP", "SNMP Time was not numeric", and "Host did not respond to SNMP".
The following is an extract from the log. Advice please. I'm not sure what they relate, and they seem contradictory.
Thanks.
02/16/2008 10:55:02 PM - CMDPHP: Poller[0] DEBUG: SQL Exec: "update host set status = '3', status_event_count = '0', status_fail_date = '2008-02-16 21:45:03', status_rec_date = '2008-02-16 22:45:03', status_last_error = 'Host did not respond to SNMP', min_time = '0.00000', max_time = '10.67000', cur_time = '0', avg_time = '0.0004199967864357', total_polls = '131634', failed_polls = '938', availability = '99.287418144248' where hostname = '127.0.0.1'"
02/16/2008 10:55:02 PM - CMDPHP: Poller[0] Host[1] NOTICE: HOST EVENT: Host Returned from DOWN State:
02/16/2008 10:55:02 PM - CMDPHP: Poller[0] Host[1] SNMP: Device does not require SNMP
02/16/2008 11:00:04 PM - POLLER: Poller[0] WARNING: Poller[0] Host[1] SNMP Time was not numeric
02/16/2008 10:55:02 PM - CMDPHP: Poller[0] DEBUG: SQL Exec: "UPDATE poller_item SET rrd_next_step = rrd_step - 300 WHERE (rrd_next_step < 0 AND host_id >= 0 AND host_id <= 3)"
--
Dan Langille - http://www.langille.org/
Dan Langille - http://www.langille.org/
solved
--
Dan Langille - http://www.langille.org/
Dan Langille - http://www.langille.org/
sure....
The settings used when the problem appeared have been changed.
Current settings are:
description: localhost
hostname: 127.0.0.1
host template: linux local machine
notes: [empty]
disabled host: [not checked]
Downed Device Detection: snmp
ping timeout value: 400
ping retry count: 1
snmp version: Version 1
snmp community: public
snmp port: 161
SNMP timeout: 500
Maximum OID's Per Get Request: 10
Current settings are:
description: localhost
hostname: 127.0.0.1
host template: linux local machine
notes: [empty]
disabled host: [not checked]
Downed Device Detection: snmp
ping timeout value: 400
ping retry count: 1
snmp version: Version 1
snmp community: public
snmp port: 161
SNMP timeout: 500
Maximum OID's Per Get Request: 10
--
Dan Langille - http://www.langille.org/
Dan Langille - http://www.langille.org/
Localhost (127.0.0.1)
SNMP Information
System: FreeBSD supernews.example.org 6.2-STABLE FreeBSD 6.2-STABLE #6: Fri Feb 9 15:54:22 PST 2007 dan@supernews.example.org:/usr/obj/usr/src/sys/OPTI amd64
Uptime: 32306022 (3 days, 17 hours, 44 minutes)
Hostname: supernews.example.org
Location: SUPERNEWS
Contact: dan@example.org
NOTE: in the above, several domain names have been changed to 'example'.
SNMP Information
System: FreeBSD supernews.example.org 6.2-STABLE FreeBSD 6.2-STABLE #6: Fri Feb 9 15:54:22 PST 2007 dan@supernews.example.org:/usr/obj/usr/src/sys/OPTI amd64
Uptime: 32306022 (3 days, 17 hours, 44 minutes)
Hostname: supernews.example.org
Location: SUPERNEWS
Contact: dan@example.org
NOTE: in the above, several domain names have been changed to 'example'.
--
Dan Langille - http://www.langille.org/
Dan Langille - http://www.langille.org/
Is this what you need?gandalf wrote:And an snmpwalk against this host using the mentioned parms would be fine
Reinhard
$ snmpwalk -v1 -c public localhost | head
SNMPv2-MIB::sysDescr.0 = STRING: FreeBSD supernews.example.org 6.2-STABLE FreeBSD 6.2-STABLE #6: Fri Feb 9 15:54:22 PST 2007 dan@supernews.exampleorg:/usr/obj/usr/src/sys/OPTI amd64
SNMPv2-MIB::sysObjectID.0 = OID: NET-SNMP-MIB::netSnmpAgentOIDs.8
DISMAN-EVENT-MIB::sysUpTimeInstance = Timeticks: (32326500) 3 days, 17:47:45.00
SNMPv2-MIB::sysContact.0 = STRING: dan@example.org
SNMPv2-MIB::sysName.0 = STRING: supernews.example.org
SNMPv2-MIB::sysLocation.0 = STRING: SUPERNEWS
SNMPv2-MIB::sysServices.0 = INTEGER: 76
SNMPv2-MIB::sysORLastChange.0 = Timeticks: (1) 0:00:00.01
SNMPv2-MIB::sysORID.1 = OID: SNMPv2-MIB::snmpMIB
SNMPv2-MIB::sysORID.2 = OID: TCP-MIB::tcpMIB
Again, domain names have been changed to 'example'.
--
Dan Langille - http://www.langille.org/
Dan Langille - http://www.langille.org/
- fmangeant
- Cacti Guru User
- Posts: 2345
- Joined: Fri Sep 19, 2003 8:36 am
- Location: Sophia-Antipolis, France
- Contact:
Hi
can you try to increase 'SNMP timeout' for that host in Cacti ?
can you try to increase 'SNMP timeout' for that host in Cacti ?
[size=84]
[color=green]HOWTOs[/color] :
[list][*][url=http://forums.cacti.net/viewtopic.php?t=15353]Install and configure the Net-SNMP agent for Unix[/url]
[*][url=http://forums.cacti.net/viewtopic.php?t=26151]Install and configure the Net-SNMP agent for Windows[/url]
[*][url=http://forums.cacti.net/viewtopic.php?t=28175]Graph multiple servers using an SNMP proxy[/url][/list]
[color=green]Templates[/color] :
[list][*][url=http://forums.cacti.net/viewtopic.php?t=15412]Multiple CPU usage for Linux[/url]
[*][url=http://forums.cacti.net/viewtopic.php?p=125152]Memory & swap usage for Unix[/url][/list][/size]
[color=green]HOWTOs[/color] :
[list][*][url=http://forums.cacti.net/viewtopic.php?t=15353]Install and configure the Net-SNMP agent for Unix[/url]
[*][url=http://forums.cacti.net/viewtopic.php?t=26151]Install and configure the Net-SNMP agent for Windows[/url]
[*][url=http://forums.cacti.net/viewtopic.php?t=28175]Graph multiple servers using an SNMP proxy[/url][/list]
[color=green]Templates[/color] :
[list][*][url=http://forums.cacti.net/viewtopic.php?t=15412]Multiple CPU usage for Linux[/url]
[*][url=http://forums.cacti.net/viewtopic.php?p=125152]Memory & swap usage for Unix[/url][/list][/size]
Sure can.fmangeant wrote:Hi
can you try to increase 'SNMP timeout' for that host in Cacti ?
What else should I do?
Please note, as detailed in the URL at "Sun Feb 17, 2008 2:19 pm Post subject: solved", the problem no longer presents itself. The situation went away after making a few adjustments.
This questions sounds sarcastic, but in a effort to track down the original cause of the problem: which adjustments do you think I should undo to attempt to reproduce the issue?
--
Dan Langille - http://www.langille.org/
Dan Langille - http://www.langille.org/
SNMP Time was not numeric
Gandalf:gandalf wrote:Thanks. Now, please post the output at the upper left corner, just under the device's name. And an snmpwalk against this host using the mentioned parms would be fine
Reinhard
Not sure if this is an issue you guys are still investigating but I'll add my own recent experience to the discussion.
I've been bringing up a new cacti server to replace the current production box.
The new server is a vm with 2 cpus and 1 gb of ram running:
SLES 9
RRDTool 1.2.27
Apache 2.2.8
NET-SNMP 5.4.1
PHP 5.2.5
MySQL 4.1.22
Cacti 0.8.7b
With everything built and the necessary workarounds in place to support the cacti url as an alias rather than subdirectory under htdocs, things have worked quite well. Only two outstanding issues:
1>. I'm unable to purge the cacti.log from within the system tools I always get the following error:
Error: Unable to clear log, no write permissions.
This is strange given that when I SU to the cactiuser on the server I can create, delete, edit, etc any file in the /log /rra directories. Not a big deal, just a curiosity.
2>. SNMP Time was not numeric messages at the start of each polling cycle. I was seeing the following error for each host target at the start of the polling cycle for the hosts that were currently configured for graphing:
03/29/2008 11:31:01 PM - POLLER: Poller[0] WARNING: Poller[0] Host[41] SNMP Time was not numeric
03/29/2008 11:31:01 PM - POLLER: Poller[0] WARNING: Poller[0] Host[42] SNMP Time was not numeric
03/29/2008 11:31:01 PM - POLLER: Poller[0] WARNING: Poller[0] Host[46] SNMP Time was not numeric
03/29/2008 11:31:01 PM - POLLER: Poller[0] WARNING: Poller[0] Host[44] SNMP Time was not numeric
03/29/2008 11:31:01 PM - POLLER: Poller[0] WARNING: Poller[0] Host[48] SNMP Time was not numeric
03/29/2008 11:31:02 PM - POLLER: Poller[0] WARNING: Poller[0] Host[52] SNMP Time was not numeric
03/29/2008 11:31:04 PM - POLLER: Poller[0] WARNING: Poller[0] Host[50] SNMP Time was not numeric
03/29/2008 11:31:07 PM - POLLER: Poller[0] WARNING: Poller[0] Host[43] SNMP Time was not numeric
03/29/2008 11:31:09 PM - POLLER: Poller[0] WARNING: Poller[0] Host[51] SNMP Time was not numeric
03/29/2008 11:31:12 PM - POLLER: Poller[0] WARNING: Poller[0] Host[47] SNMP Time was not numeric
03/29/2008 11:31:16 PM - POLLER: Poller[0] WARNING: Poller[0] Host[49] SNMP Time was not numeric
03/29/2008 11:31:18 PM - POLLER: Poller[0] WARNING: Poller[0] Host[45] SNMP Time was not numeric
03/29/2008 11:31:28 PM - SYSTEM STATS: Time:27.4646 Method:cmd.php Processes:8 Threads:N/A Hosts:13 HostsPerProcess:2 DataSources:1168 RRDsProcessed:584
The hosts/graphs in question are configured for 1 minute polling using 64-bit counters and snmpv 3 with an snmp timeout of 500 miliseconds.
Interestingly enough the error message appeared to be somewhat cosmetic. Despite the error message, the rrds and graphs were being updated every 60 seconds.
As I continued to populate the system with new devices and graphs for the switch ports for the devices, the messages continued but my graphs continued to work as expected with no gaps etc. Turning on debug never yielded any indication of an error etc.
The number of graph targets in this new system is significantly higher than in the old system and looked like it would take cmd.php more than 60 seconds to process all of the graphs. I went ahead and installed Spine 0.8.7a. Its working great.
Check out the difference in the polling times:
03/29/2008 11:39:25 PM - SYSTEM STATS: Time:24.5518 Method:cmd.php Processes:8 Threads:N/A Hosts:13 HostsPerProcess:2 DataSources:1168 RRDsProcessed:584
03/29/2008 11:40:03 PM - SYSTEM STATS: Time:2.2443 Method:spine Processes:8 Threads:6 Hosts:13 HostsPerProcess:2 DataSources:1168 RRDsProcessed:584
The most interesting and unexpected change was that the SNMP Time was not numeric messages have vanished. I'm not sure if this is to be expected but I thought I would share it with you if it helps identify the cause of the SNMP Time error for others.
Doug
- gandalf
- Developer
- Posts: 22383
- Joined: Thu Dec 02, 2004 2:46 am
- Location: Muenster, Germany
- Contact:
Re: SNMP Time was not numeric
Write permission must be granted to the httpd's user as well to delete logs using the web screenshosierd wrote:1>. I'm unable to purge the cacti.log from within the system tools I always get the following error:
Error: Unable to clear log, no write permissions.
This is strange given that when I SU to the cactiuser on the server I can create, delete, edit, etc any file in the /log /rra directories. Not a big deal, just a curiosity.
Reinhard
Hello,
I'm doing a training period in IT domain and my work is to supervise a network.I make the choice with Nagios + Cacti.
So I install cacti 0.8.7b and I have got the same problem :
04/30/2008 03:00:01 PM - POLLER: Poller[0] WARNING: Poller[0] Host[9] SNMP Time was not numeric.
Then I think I found a bug in the PHP code in the "ping.php" file (/var/www/cacti/lib/ping.php).I'm not a pro in PHP but I let developers confirm this bug.
Then, in the ping.php file, we have :
/* snmp test */
if (($avail_method == AVAIL_SNMP) || (($avail_method == AVAIL_SNMP_AND_PING) && ($ping_result == true))) {
if ($this->host["snmp_community"] != "") {
$snmp_result = $this->ping_snmp();
}else{
$snmp_result = true;
}
must be change in ...
/* snmp test */
if (($avail_method == AVAIL_SNMP) || (($avail_method == AVAIL_SNMP_AND_PING) && ($ping_result == true))) {
$snmp_result = $this->ping_snmp();
}
because if we put community to "" , the variable $ping->snmp_status stay to "down" and the variable is not update.
Then, that makes an error in cacti log (poller warning ...).
I test it and the error disappears!
Can developer confirm that bug ?
(Sorry if it's not a bug )
I'm doing a training period in IT domain and my work is to supervise a network.I make the choice with Nagios + Cacti.
So I install cacti 0.8.7b and I have got the same problem :
04/30/2008 03:00:01 PM - POLLER: Poller[0] WARNING: Poller[0] Host[9] SNMP Time was not numeric.
Then I think I found a bug in the PHP code in the "ping.php" file (/var/www/cacti/lib/ping.php).I'm not a pro in PHP but I let developers confirm this bug.
Then, in the ping.php file, we have :
/* snmp test */
if (($avail_method == AVAIL_SNMP) || (($avail_method == AVAIL_SNMP_AND_PING) && ($ping_result == true))) {
if ($this->host["snmp_community"] != "") {
$snmp_result = $this->ping_snmp();
}else{
$snmp_result = true;
}
must be change in ...
/* snmp test */
if (($avail_method == AVAIL_SNMP) || (($avail_method == AVAIL_SNMP_AND_PING) && ($ping_result == true))) {
$snmp_result = $this->ping_snmp();
}
because if we put community to "" , the variable $ping->snmp_status stay to "down" and the variable is not update.
Then, that makes an error in cacti log (poller warning ...).
I test it and the error disappears!
Can developer confirm that bug ?
(Sorry if it's not a bug )
I am using the version 3 of SNMP and I have got no snmp community.
You don't have to put a snmp community in the version 3, if I learn my lessons correctly.
Because you need to use the username only.
Furthermore, you can't put it because there is no "snmp community" field when you choose "Version3" in the SNMP Option.
Then that generate an error in the cacti log file which say "SNMP Time was not numeric" because ping->snmp_status is always on "down" (As I know,"down" is not numeric ).
You don't have to put a snmp community in the version 3, if I learn my lessons correctly.
Because you need to use the username only.
Furthermore, you can't put it because there is no "snmp community" field when you choose "Version3" in the SNMP Option.
Then that generate an error in the cacti log file which say "SNMP Time was not numeric" because ping->snmp_status is always on "down" (As I know,"down" is not numeric ).
Who is online
Users browsing this forum: No registered users and 3 guests