- observium - lists.observium.org

A couple of questions on port details
by Aaron Mayfield 16 Nov '12

16 Nov '12

1. I notice that with the Real time graph running on a port, quite often the graph will calculate a negative number resulting in gaps in the graphs. Has anyone else experienced this? Is it expected? 2. The Arp Table output for my ports are blank. Does this feature work for you? If so, does anyone have this working with Arista switches?

5 8

Suggestion: All Ports
by Mattias Gyllenvarg 16 Nov '12

16 Nov '12

Hi As our installation is getting heavy and I have seen more and more use for the "Ports/All Ports" page. I becomes clear that the page should not load anything but the search bar until you actually klick the search button. Saves me some minutes of waiting everyday. -- *Med Vänliga Hälsningar - Best Regards* *Mattias Gyllenvarg* *Nätutveckling* Bredband2 Tel: +46 406219712

5 11

Alerting Contrib?
by Storer, Darren 15 Nov '12

15 Nov '12

Hi, I recently had a very good experience setting up Observium with in a mixed switch, router and server environment; everything was so much slicker than Nagios. Despite the inclusion of "Alerting" tabs, I was assured on the IRC channel that there is no alerting functionality with Obs at this time. Has anyone produced their own addon/scripts/strategy for basic alerting? Best regards Darren

4 6

6500 Fabric Ingress/Egress scale
by Robert Williams 11 Nov '12

11 Nov '12

Hi Guys, Firstly, apologies for the screen-shots / attachments, but they are required to demonstrate the oddness... I think there is a scaling issue on (or I'm not understanding) the Fabric load bars for the Catalyst 6k. In short, the load bars (to the right of the mini-graph) always seem to show zero, or near zero, even though load is not zero. Example below or attached (or missing, depending on how this list handles images :)) [cid:image003.png@01CDBF49.C8C03550] All bars are showing 2% Ingress and 1% Egress. Below is a zoomed in version of the graphic for Slot 1 Fabric 0: [cid:image002.png@01CDBF48.A37B26C0] This shows ingress is 4% and egress is 2% right now. Load is quite low at the moment so it's not an excellent example I agree - but I can honestly say that even when it's high the most I see on those bars is 2%. So, am I being stupid (quite possible) and these bars mean something else or is there some sort of issue with them or their maths? We have 11 of these chassis, and I've checked them all. They all either show zero or are reading 1 or 2% despite having much higher real loads. Such as this one: [cid:image001.png@01CDBF49.E3BFD450] Maybe if anyone else can check their bars and let me know if it's just me? Cheers! Robert Williams Custodian DataCentre email: Robert(a)CustodianDC.com

4 9

Important: alert mails disabled by default
by Tom Laermans 10 Nov '12

10 Nov '12

Hi, This is a notification about an important change! In some cases, Observium sends out alert emails upon events. While we always advised to disable these as they are mostly rubbish, they were, until now, enabled by default. They are now disabled, since revision 3406. This means that when you update you will no longer receive the alert emails. If you want to keep receiving these alerts after you svn up to r3406 or higher, add this to your config.php: $config['alerts']['email']['enable'] = TRUE Hopefully we'll have a decent threshold/event/alert system soon; ideas about this are very welcome in our IRC channel. Tom

1 0

Maps.php
by Dermot Williams 09 Nov '12

09 Nov '12

Hi, Quick question (the answer to which is possibly 'nothing') but - what is map.php supposed to do? - Dermot

5 15

scaling towards >100GE LAGs
by Job Snijders 09 Nov '12

09 Nov '12

Hi all, I was facing some issues with observium and 16x10GE LAGs, where RRD would flip the finger on me. I propose we let the RRDs accept up to a terabit/sec instead of the current 100gigabit/sec limit. In a year or so we can up it again. :-) (or make it a configurable option!) BTW: If you currently have ports which go over 100GE you need to use 'rrdtool dump' -> edit xml file and change "1.2500000000e+10" to "1.2500000000e+11" and then use 'rrdtool restore' to get the XML back into RRD format. root@observium:/opt/observium# diff includes/polling/ports.inc.php.orig includes/polling/ports.inc.php 339,353c339,353 < DS:INOCTETS:DERIVE:600:0:12500000000 \ < DS:OUTOCTETS:DERIVE:600:0:12500000000 \ < DS:INERRORS:DERIVE:600:0:12500000000 \ < DS:OUTERRORS:DERIVE:600:0:12500000000 \ < DS:INUCASTPKTS:DERIVE:600:0:12500000000 \ < DS:OUTUCASTPKTS:DERIVE:600:0:12500000000 \ < DS:INNUCASTPKTS:DERIVE:600:0:12500000000 \ < DS:OUTNUCASTPKTS:DERIVE:600:0:12500000000 \ < DS:INDISCARDS:DERIVE:600:0:12500000000 \ < DS:OUTDISCARDS:DERIVE:600:0:12500000000 \ < DS:INUNKNOWNPROTOS:DERIVE:600:0:12500000000 \ < DS:INBROADCASTPKTS:DERIVE:600:0:12500000000 \ < DS:OUTBROADCASTPKTS:DERIVE:600:0:12500000000 \ < DS:INMULTICASTPKTS:DERIVE:600:0:12500000000 \ < DS:OUTMULTICASTPKTS:DERIVE:600:0:12500000000 ".$config['rrd_rra']); --- > DS:INOCTETS:DERIVE:600:0:125000000000 \ > DS:OUTOCTETS:DERIVE:600:0:125000000000 \ > DS:INERRORS:DERIVE:600:0:125000000000 \ > DS:OUTERRORS:DERIVE:600:0:125000000000 \ > DS:INUCASTPKTS:DERIVE:600:0:125000000000 \ > DS:OUTUCASTPKTS:DERIVE:600:0:125000000000 \ > DS:INNUCASTPKTS:DERIVE:600:0:125000000000 \ > DS:OUTNUCASTPKTS:DERIVE:600:0:125000000000 \ > DS:INDISCARDS:DERIVE:600:0:125000000000 \ > DS:OUTDISCARDS:DERIVE:600:0:125000000000 \ > DS:INUNKNOWNPROTOS:DERIVE:600:0:125000000000 \ > DS:INBROADCASTPKTS:DERIVE:600:0:125000000000 \ > DS:OUTBROADCASTPKTS:DERIVE:600:0:125000000000 \ > DS:INMULTICASTPKTS:DERIVE:600:0:125000000000 \ > DS:OUTMULTICASTPKTS:DERIVE:600:0:125000000000 ".$config['rrd_rra']); root@observium:/opt/observium# Kind regards, Job Snijders -- AS5580 - Atrato IP Networks

2 2

Re: [Observium] observium Digest, Vol 28, Issue 7
by Joshua Lebo 08 Nov '12

08 Nov '12

77 Devices with 2686 ports. > On Nov 7, 2012, at 1:07 AM, Morgan McLean wrote: > > > Just curious what the average size of an observium installation is among > the users on the email list. > > > > I have ~150 devices with ~6000 ports across three sites. Anybody else? > > > > Morgan > > _______________________________________________ > > observium mailing list > > observium(a)observium.org > > http://postman.memetic.org/cgi-bin/mailman/listinfo/observium > >

2 1

Re: [Observium] Out of curiosity, how big is everyones installation?
by Steve Costaras 07 Nov '12

07 Nov '12

would be possible, but would much rather have that as a php or other plugin to the gui so it can happen all on the same screen/time of object creation/deletion. It avoids having people who do NOT have unix skills needing to get into the os for simple day/day operation (which here is adding/removing devices). This though is secondary to grouping for large deployments. I'm only dealing with one small site currently, but we numerous soho and other locations around the world and ancillary sites/DC's. using dns name doesn't work when there are hundreds/thousands of devices of the same types (takes /way/ to long to find what you're looking for or to focus on only certain logical groupings of devices. community strings don't work as there is no central authority to control what goes into that field (not all devices are controlled by us, but we're tasked to monitor/trend/troublehsoot). >-----Original Message----- >From: Tom Laermans [mailto:tom.laermans@powersource.cx] >Sent: Wednesday, November 7, 2012 07:07 AM >To: observium(a)observium.org >Subject: Re: [Observium] Out of curiosity, how big is everyones installation? > >Indeed - smokeping integration in Observium will never(*) write into >other config files; just like we won't add devices to rancid, etc. >Scripting against mysql -e "SELECT hostname FROM devices" output is easy >though. > >(*) right to change reserved > >On 7/11/2012 14:04, Jason Lixfeld wrote: >> Why not write a script to pull the hostnames out of the database at some internal into the smokeping config file? >> >> -- >> >> Sent from my mobile device >> >> >> On 2012-11-07, at 8:00 AM, "Steve Costaras" <stevecs(a)chaven.com> wrote: >> >>>> -----Original Message----- >>>> From: Mattias Gyllenvarg [mailto:Mattias.Gyllenvarg@Bredband2.se] >>>> Only our Core/Dist devices are in Observium >>>> >>>> Devices 183 ( 177 up 1 down 2 ignored 3 disabled ) >>>> Ports 9019 ( 4222 up 179 down 542 ignored 3957 shutdown ) >>>> >>>> Access devices (1500 more) may be added if there had been a good way too >>>> categorize. >>> Still testing/playing here with only 80 devices 4077 ports, but the above sentiments are shared, we have several thousand more devices as well, but without a grouping/categorization system as well as some tighter integration w/ smokeping (so by adding the device in the gui it creates/adds it in the smokeping configuration) don't see this really continuing. >>> >>> Also starting to look at Reconnoiter and other tools. >>> >>> >>> >>> >>> _______________________________________________ >>> observium mailing list >>> observium(a)observium.org >>> http://postman.memetic.org/cgi-bin/mailman/listinfo/observium >> _______________________________________________ >> observium mailing list >> observium(a)observium.org >> http://postman.memetic.org/cgi-bin/mailman/listinfo/observium > >_______________________________________________ >observium mailing list >observium(a)observium.org >http://postman.memetic.org/cgi-bin/mailman/listinfo/observium >

1 0

Re: [Observium] Out of curiosity, how big is everyones installation?
by Steve Costaras 07 Nov '12

07 Nov '12

>-----Original Message----- >From: Mattias Gyllenvarg [mailto:Mattias.Gyllenvarg@Bredband2.se] >Only our Core/Dist devices are in Observium > >Devices 183 ( 177 up 1 down 2 ignored 3 disabled ) >Ports 9019 ( 4222 up 179 down 542 ignored 3957 shutdown ) > >Access devices (1500 more) may be added if there had been a good way too >categorize. > Still testing/playing here with only 80 devices 4077 ports, but the above sentiments are shared, we have several thousand more devices as well, but without a grouping/categorization system as well as some tighter integration w/ smokeping (so by adding the device in the gui it creates/adds it in the smokeping configuration) don't see this really continuing. Also starting to look at Reconnoiter and other tools.

3 2

Re: [Observium] observium Digest, Vol 28, Issue 5
by Dan Nelson 07 Nov '12

07 Nov '12

1 Virtual Server, 67 Devices, 1265 Ports running great. Thanks Dan Nelson Nutraceutical Corporation Network Administrator 801-334-3702

1 0

Procedure to add a MIB file for a device already monitored by Observium
by Charles Nadeau 01 Nov '12

01 Nov '12

Hello, I put a MIB file in /opt/observium/mibs/ for a device whose interfaces I already monitor with SNMP. In order to have Observium uses the MIB to monitor all its components do I just run "/opt/observium/discovery.php -h new" again or do I have to modify "config.php" then run "/opt/observium/discovery.php -h new"? Thanks! Charles -- Charles Nadeau Ph.D. http://charlesnadeau.blogspot.com/

3 2

Re: [Observium] observium Digest, Vol 28, Issue 1
by Charles Nadeau 01 Nov '12

01 Nov '12

On Thu, Nov 1, 2012 at 8:00 PM, <observium-request(a)observium.org> wrote: > Date: Thu, 01 Nov 2012 12:43:25 +0400 > From: Nikolay Shopik <shopik(a)inblock.ru> > To: Observium Network Observation System <observium(a)observium.org> > Subject: Re: [Observium] Procedure to add a MIB file for a device > already monitored by Observium > Message-ID: <5092362D.9000403(a)inblock.ru> > Content-Type: text/plain; charset=UTF-8 > > Nope, > > MIB just describe these counters it won't add support to observium. > Patch is need to add support of new device/mib Can you tell me where and what to add, I could give it a go then share it if I am successful. Thanks! Charles

2 1

IP SLA
by John Macleod 30 Oct '12

30 Oct '12

I have setup two devices with IP SLA, however I am not getting anything in Observium. One is a Cisco 877W 12.4(24)T6, the other is a Cisco 2901 15.1(4)M3. I verified that the IP SLA module is enabled, I must be missing something?

6 16

fresh installation, no login screen
by Elad Eyal (Ubiqam) 29 Oct '12

29 Oct '12

I've followed the installation instructions to the letter, and while I can get to the observium webpage, I can see observium logo and the text "Big Graphs | Widescreen | Not logged in!" on the top right of the page, I can't see anything else. Furthermore, when I look at the HTML source of the page delivered, it ends with " Not logged in!" and not with </HTML> as I would expect. The machine is a brand new Amazon EC3 VM running Ubuntu 12.04.1 LTS. I'd appriciate any help. I don't see anything of interest in the apache logs. The same client (browser) works OK for the live demo. thanks Elad

4 4

Custom port descriptinos
by Michael Cushard 26 Oct '12

26 Oct '12

Hi, I see that the ability to add custom port descriptins is available, but I cannot find any instructions to how to actually do it. What is the procedure for adding custom port descriptions? Regards, Michael Cushard

1 0

Device Traffic Graph
by Bruno Galeppe 25 Oct '12

25 Oct '12

Hi, I set up several devices on Obervium, no problem so far, except for the Traffic Graphs which remain empty. For example : http://img4.hostingpics.net/pics/168961data.jpg The RRD file exist, but the data always return 0.00 How can I debug that ? Is there an snmpwalk command to help understand the problem ? Regards,

4 21

Various questions (BGP/Discovery/Pseudowires/Hostname truncating)
by Jason Lixfeld 22 Oct '12

22 Oct '12

First off, let me say thanks to the developers of Observium. This piece of code is one of the most amazing pieces of code I have used for my network monitoring purposes to date. I have a few questions that I'm hoping someone can answer - BGP: ---- I have a few eBGP speakers that live inside a VRF. Observium picks up the VRFs just fine, but the routing tab for the device only seems to show the BGP sessions that exist in the global (non-VRF) table. Is it possible to show the BGP sessions that exist inside a VRF? Pseudowires: ------------ I see lots of data inside the pseudowires table on the observium database, but I don't see any of that inside the pseudowires tab from within a device where I know pseudowires are configured. These pseudowires exist on IOS boxes, IOS-XR boxes and IOS-XE boxes. I do see a few. Observium > Ports > Pseudowires shows me 11 pseuowires. 6 of them are duplicates of one another while 5 are unique. The database shows hundreds. While looking through a ./discovery.php -d I ran against one of our XR boxes, I did see this: Cisco Pseudowires : SQL[SELECT * FROM `pseudowires` WHERE `device_id` = '1'] DEBUG: SNMP Auth options = -v2c -c ************ /usr/bin/snmpbulkwalk -v2c -c ************ -OQUs -m CISCO-IETF-PW-MPLS-MIB -M /opt/observium/mibs udp:router:161 cpwVcID cpwVcID = No Such Instance currently exists at this OID DEBUG: SNMP Auth options = -v2c -c ************ /usr/bin/snmpbulkwalk -v2c -c ************ -OQUs -m CISCO-IETF-PW-MPLS-MIB -M /opt/observium/mibs udp:router:161 cpwVcName cpwVcName = No Such Instance currently exists at this OID DEBUG: SNMP Auth options = -v2c -c ************ /usr/bin/snmpbulkwalk -v2c -c ************ -OQUs -m CISCO-IETF-PW-MPLS-MIB -M /opt/observium/mibs udp:router:161 cpwVcType cpwVcType = No Such Instance currently exists at this OID DEBUG: SNMP Auth options = -v2c -c ************ /usr/bin/snmpbulkwalk -v2c -c ************ -OQUs -m CISCO-IETF-PW-MPLS-MIB -M /opt/observium/mibs udp:router:161 cpwVcPsnType cpwVcPsnType = No Such Instance currently exists at this OID DEBUG: SNMP Auth options = -v2c -c ************ /usr/bin/snmpbulkwalk -v2c -c ************ -OQUs -m CISCO-IETF-PW-MPLS-MIB -M /opt/observium/mibs udp:router:161 cpwVcDescr cpwVcDescr = No Such Instance currently exists at this OID DEBUG: SNMP Auth options = -v2c -c ************ /usr/bin/snmpbulkwalk -v2c -c ************ -OQUs -m CISCO-IETF-PW-MPLS-MIB -M /opt/observium/mibs udp:router:161 cpwVcMplsPeerLdpID cpwVcMplsPeerLdpID = No Such Instance currently exists at this OID That seems to show that pseudowire support for XR isn't there yet? Is that on the road map, by chance? Discovery: ---------- There are a couple of devices that aren't getting discovered for some reason. The IP that their hostname resolves to is within the 'nets' arrays configured inside config.php and I have other SNMP devices that reach those boxes, so I know the SNMP side is sound. Likewise, I can see the device in question in the CDP table of an adjacent device, so I know that part is working as well. I've looked at a ./discovery -d of an adjacent device hoping to see a clue about why the device is being skipped, but nothing is striking out at me. Hostname truncating: -------------------- We name our devices <device><device_number>.<pop>.<city>.domain.tld. In the devices view, it shows the full device name no problem. In the VRF view, for example, however, it shows only <device><device_number>, truncating the pop and city portions. If anyone has any insight or guidance on these issues, I'd be grateful! Thanks in advance.

2 3

Unable to get non-MySQL Auth working - 500 error
by Chris Moody 19 Oct '12

19 Oct '12

Hello. First off, let me say that this project is AWESOME!!! I am blown away at how comprehensive the graphing is across our network gear. Bravo. That said, I need a hand in figuring out why I cannot seem to get any auth-method other than mysql to work. If I so much as toggle to any other auth-method (ldap or http-auth), I immediately get a 500 'Internal Error' when I try to hit the Observium login page. I'm ultimately trying to get auth to the application to be handled via LDAP, but I can do it via either the observium-ldap component or via http-auth combined with mod_authz_ldap... makes me no difference. I've had some success w. mod_authz_ldap in that I can get authentication to actually pass for a user account (I am able to bypass the login promptsuccessfully), but as soon as I auth, I hit the dreaded 500. I'm not able to find anything in any http logs regarding what might be possibly throwing the 500. lsof lists any open filehandles for httpd, so I've checked every possible file that logs could be directed to, but no cluesthus far. This particular box is a RHEL5 running Observium 0.12.10.3396 Any ideas are welcome. Cheers, -Chris

2 1

unix agent polling broken in 3397 - likely a couple of versions before too
by Chris Stone 18 Oct '12

18 Oct '12

No data for servers using the unix agent for a couple of weeks now. Have Observium build 3397 installed. Just had a chance to do some troubleshooting... Running poller update on a specific server in debug mode reveals the apparent problem: ======================================================================================= including: includes/polling/unix_agent.inc.php PHP Warning: include(includes/polling/unix_agent.inc.php): failed to open stream: No such file or directory in /opt/observium/includes/polling/functions.inc.php on line 192 Warning: include(includes/polling/unix_agent.inc.php): failed to open stream: No such file or directory in /opt/observium/includes/polling/functions.inc.php on line 192 PHP Warning: include(): Failed opening 'includes/polling/unix_agent.inc.php' for inclusion (include_path='.:/usr/share/pear:/usr/share/php') in /opt/observium/includes/polling/functions.inc.php on line 192 Warning: include(): Failed opening 'includes/polling/unix_agent.inc.php' for inclusion (include_path='.:/usr/share/pear:/usr/share/php') in /opt/observium/includes/polling/functions.inc.php on line 192 ======================================================================================= Fixed (worked around) by running: ln -s /opt/observium/includes/polling/unix-agent.inc.php /opt/observium/includes/polling/unix_agent.inc.php Added to Jira as well: http://jira.observium.org/browse/OBSERVIUM-154 Chris -- Chris Stone AxisInternet, Inc. www.axint.net

2 4

Backing Up Observium
by Mahomed Hussein 18 Oct '12

18 Oct '12

Hi I need to ensure our install of Observium is fully backed up and I'd rather not backup the entire VM. Could you please confirm if all I need to backup are: * /opt/observium directory * Observium database * /etc/apache2/sites-enabled/000-default What I'd like to achieve from the backups is to spin up a new VM instance, run through the install steps to get the required components installed, then restore the files so that Observium then just runs as it was at the point of the last backup. I have tried checking the FAQ and a search of the mailing list but I couldn't find any specific instructions. Any help will be greatly appreciated. Thanks in advance. Kind regards, Mahomed Hussein Custodian DataCentre email: Mahomed(a)CustodianDC.com

10 23

Port Utilization Alerts
by Adam Blackington 18 Oct '12

18 Oct '12

Observium folk, Anyway I can configure alerting based on utilization? For example, if I want to receive an email if a port hits 75% port utilization. --Adam Blackington

3 2

HIgh CPU on ASR
by Roza 17 Oct '12

17 Oct '12

After adding ASR 1002 (used as PPPOE termination) to Observium, CPU on ASR raised to 90%, and stayed that way, untill I blocked it (after 30 minutes). Normal CPU is aroiund 10%. Is this a normal? Don't think so.... ASR1002 has cca 5000 connected users. iRoza

4 4

Cisco CUCM Support
by John Macleod 13 Oct '12

13 Oct '12

Has anyone looked into fully supporting the Cisco UCM stuff? It looks like it is mainly a mix of Cisco CCM and Cisco/HP/IBM hardware mibs depending on what platform it is on. I specifically have UCM5xx series, the SMB PBX in a box - CME variant and the larger full Unified Communications Manager 5000 systems, running the linux-derived UCM apps on Cisco branded UCS, IBM or HP hardware. I did find this: http://smbitsolutions.wordpress.com/2011/05/02/2-may-2011-1341/ I do have a system I could expose for testing if someone is up for this? I have the same request for Cisco CUBE (SIP-SIP) gateway support, which looks like it is just a mib on the relevant supported hardware. I believe it is the same mib for standard SIP monitoring on other Cisco platforms, but I could be wrong. At the moment the UCM 5000 shows as a generic server, with standard net-snmp type support, and the CUBE devices just show as routers. I currently use a Cacti install alongside my Observium systems for the sole purpose of monitoring the above. Would be nice to stick to one. Thanks, John

1 0

Observium behind proxy
by Max León 10 Oct '12

10 Oct '12

Hello, I need to publish an observium instance to the DMZ. The simplest approach is by proxy using Nginx. I have the configuration set, well I think, problem is that on authentication it loops on the login, any idea why this might be? Thank you. Max Atentamente, Max León

3 3