- observium - lists.observium.org

"Device rebooted" message every 5 minutes after customer server replaced
by Andrew Beals 28 Jun '12

28 Jun '12

I upgraded one of our client's servers a couple of days ago by pushing all of its data onto a new image on AWS. Since I re-enabled Observium monitoring on it (same IP address as the old one), I've been getting a "device rebooted" message every 5 minutes from Observium. Observium :: Network Observation and Monitoring Device Rebooted : hostname 1 day, 18h 17m 23s ago. How do I make it stop? -- Andrew Beals Cell: 913-787-7538 Desk: 913-948-8782 E-mail: ABeals(a)mersoft.com

4 3

PowerDNS
by Aaron Mayfield 27 Jun '12

27 Jun '12

I think during one of the recent updates, my powerDNS graphs broke. Has anyone else experienced this? Here is the output of ./poller.php -h 8 -d (with host 8 being one of my DNS servers): **************************************************** Notice: Undefined index: poll_applications in /opt/observium/includes/polling/functions.inc.php on line 189 including: includes/polling/applications.inc.php SELECT * FROM `applications` WHERE `device_id` = '8' SQL[SELECT * FROM `applications` WHERE `device_id` = '8'] Applications: SQL[SELECT `graph` FROM `device_graphs` WHERE `device_id` = '8'] uptime netstat_ip netstat_ip_frag netstat_tcp netstat_udp netstat_icmp netstat_icmp_info netstat_snmp netstat_snmp_pkt hr_processes hr_users ucd_cpu ucd_swap_io ucd_io ucd_contexts ucd_interrupts ucd_memory ucd_load ipsystemstats_ipv4 ipsystemstats_ipv4_frag ipsystemstats_ipv6 ipsystemstats_ipv6_frag Polled in 0.938 second **************************************************** Here is the mysql query result on this: mysql> SELECT * FROM `applications` WHERE `device_id` = '8'\G *************************** 1. row *************************** app_id: 3 device_id: 8 app_type: powerdns app_state: UNKNOWN app_status: 1 row in set (0.00 sec) mysql> **************************************************** As you can see, Observium doesn't seem to be picking up on the fact that there is a registered application for the host. Does anyone have any ideas on how to troubleshoot this?

8 19

trouble ignoring switch ports - update broke functionality
by Henrik Lund Kramshøj 26 Jun '12

26 Jun '12

Hi All I have Observium running for some time and currently at [hlk@wiseguy01 observium]$ sudo svn up At revision 3310. After Observium SVN upgrade I cannot seem to ignore certain ports from Dell PowerConnect switches (might not be only those) from high port numbers I get "Device record unchanged. No update necessary." this happens currently on two switch stacks that are quite large, both 10 units single port 9/0/35 on one stack lots of ports from units 8, 9, 10 on the other stack Those cannot be ignored. I tested using ports 1/0/4 and 7/0/51 that I can ignore/unignore another port on this switch. also tried removing the switch completely and adding it again. Any pointers or ideas? Where to look for this - used to work fine. BTW I noticed to other ports were changed today - we moved a server, and changing those on a Dell 8024F gave the expected " 2 Device record updated." and were updated fine. Best regards Henrik -- Henrik Lund Kramshøj, Follower of the Great Way of Unix hlk(a)kramse.org hlk(a)solidonetworks.com +45 2026 6000 cand.scient CISSP http://solidonetworks.com/ Network Security is a business enabler

2 2

Sensors not picking up, get rediscovered 4 times a day
by Morgan McLean 26 Jun '12

26 Jun '12

My logs look like this...any idea why its not picking up? I have about 50 juniper devices, an APC device and a couple others and it doesn't seem to monitor or graph any of the sensors. I have seen some information about sensor configuration under config.php, but it wasn't very helpful. Every time the discovery runs, or maybe its the poller, not sure, it finds all these sensors for every device, but it does nothing with them. Is there a switch I need to flip? Thanks! Morgan 21/Jun/12 12:39:04Sensor Added: temperature junos 6.1.0.0 TFEB MX8021/Jun/12 12:39:04Sensor Added: temperature junos 6.1.1.0 TFEB Intake21/Jun/12 12:39:04Sensor Added: temperature junos 9.1.0.0 Routing Engine21/Jun/12 12:39:03Processor added: type junos index 6.1.0.0 descr TFEB MX8021/Jun/12 12:39:03Processor added: type junos index 7.1.0.0 descr FPC @ 0/*/*21/Jun/12 12:39:03Processor added: type junos index 7.2.0.0 descr FPC @ 1/*/*21/Jun/12 12:39:03Processor added: type junos index 9.1.0.0 descr Routing Engine

2 4

up / down gives no warning
by Ian Armstrong 24 Jun '12

24 Jun '12

Hi, Observium has been working very well over the past few months and now I want to begin tailoring it for our network specifically. We use an Ubuntu install as per the instructions and had no install problems. On the overview page we get an orange warning for a non-ignored link that goes into the (Cisco) down/down state BUT NOT when they are up/down. Is this a setting, bug or feature? If it can be changed does anyone know how - I would like a warning for any link that goes down/down and up/down. I have looked through the documentation, options and settings but find nothing. I also tried [alert][port][ifdown] as a long shot. Any help would be great. Thanks Ian

2 2

custom scripts
by mhoffmann＠arcor.de 22 Jun '12

22 Jun '12

Hello, i have many Allied Telesis Switches. Most Values are covered by the RADLAN-Mib, but some Values like Temp or PowerConsumption are not available inside SNMP. Is there a Wrapper or a other possibility to get this values? I think about a script which read the values and offer it to observium.

2 1

Apache authentication
by Ciro Iriarte 20 Jun '12

20 Jun '12

Hi, can someone confirm me if observium supports external (apache2) authentication?, I'm looking for kerberos authentication against AD. Regards, -- Ciro Iriarte http://cyruspy.wordpress.com --

4 9

How to monitor ASA connection table ?
by inlee 13 Jun '12

13 Jun '12

Hi I just installed observium. ASA is polling all the information from ASA 55xx except connection count. How can i add polling ASA connection count?

2 1

Speaking of Observium's pretty graphsŠ
by Andrew Beals 12 Jun '12

12 Jun '12

One of our clients uses a very large HP ProCurve 4108gl switch as the central hub of their network. When I'm looking at traffic graphs from it for large spaces of time, I invariably see it pegged to +/- 20Gbps because of some traffic spikes during the day (e.g. Windows clients talking to WSUS and downloading their patches). If I zoom in on one of the "empty" stretches of time, I see that there is actually some traffic flowing – it's just so small in comparison that it gets flatlined by the spikes. What I'd like to be able to do is to change the graph from a linear one in the Y (bps) axis to a logarithmic one, so I can see my traffic. For example, here's the past 24 hours: [cid:5CBC68CF-394E-403F-9F59-AB51C6450F66] (The rrdtool command that produces this graph is truly impressive and only crawls out of its blue box with the last set of CDEF statements before the giant list of AREA statements.) (All graphs generated for this e-mail use the same time span on the same hardware) I know, I know, "Use the Source, Luke!", but I've only watched the commits fly past, not actually dug into the code past the config file (and this doesn't appear to be a configuration option). rrdtool_graph() is called in three places, and fortunately only defined in one. Applying the obvious small patch: Index: html/includes/graphs/graph.inc.php =================================================================== --- html/includes/graphs/graph.inc.php (revision 3268) +++ html/includes/graphs/graph.inc.php (working copy) @@ -146,6 +146,7 @@ if ($rrd_options) { + $rrd_options .= " --logarithmic"; rrdtool_graph($graphfile, $rrd_options); if ($debug) { echo($rrd_cmd); } if (is_file($graphfile)) Nets me this: [cid:806F0C8F-84AE-4C4F-8D2D-04802B3A5EB7] Which could probably use a little more tweaking. I'd rather see SI "-bi-" units on the left rather than scientific notation. Also, it does tend to whack the CPU pretty hard figuring it all out, pegging both of the CPUs in the old ml350 I'm using for monitoring. Running "rrdgraph" at "nice +4" (or better) would likely be a win w/r/t data collection. This patch: Index: html/includes/graphs/graph.inc.php =================================================================== --- html/includes/graphs/graph.inc.php (revision 3268) +++ html/includes/graphs/graph.inc.php (working copy) @@ -146,6 +146,7 @@ if ($rrd_options) { + $rrd_options .= " --logarithmic --units=si"; rrdtool_graph($graphfile, $rrd_options); if ($debug) { echo($rrd_cmd); } if (is_file($graphfile)) Nets this: [cid:4C488D23-BF57-49B9-BBBE-C2D43F0E17AF] I'm pretty happy with this, overall. For those who are curious, the ProCurve Switch 4108gl has a claimed capacity of: routing/switching capacity 36.6 Gbps Throughput up to 71.4 mpps Being able to tweak individual graphs to show logarithmically would probably be nice, just to save some CPU for graphs that don't really need it. -- Andrew Beals Cell: 913-787-7538 Desk: 913-948-8782 E-mail: ABeals(a)mersoft.com<mailto:ABeals@mersoft.com>

2 1

Vyatta
by Adam Blackington 10 Jun '12

10 Jun '12

Any ideas on getting Vyatta recognized? We keep seeing it as a generic device, and can't seem to get it working. :-/ Running v2c communities. *Adam Blackington* | Solutions Engineer *LionLink* *Networks *| www.lionlink.net | AS54098 O: (571) 451.2300 | D: (571) 482.1490 | E: ablackington(a)lionlink.net

2 17