- observium - lists.observium.org

Re: [Observium] Syslog msg empty temporary fix
by Robert Williams 27 Aug '14

27 Aug '14

Hi, Further to Pav's comments earlier, I’ve found that setting the “no-parse” flag in syslog-ng stops it from messing with the string and restores some reasonable sanity to the messages which get passed to Observium. Clearly something has changed within syslog-ng from Ubuntu 12->14 and this new issue is nothing to do with the Observium-importing-the-message element. For anyone suffering the same fate, the actual setting syntax to be used within the Observium definition for syslog-ng is: source s_net { udp(flags(no-parse)); }; This restores 100% normal message structure for /most/ of the devices I’ve just tested with, including all the IOS 15.x ones which had all started showing simply "%" as the message content. The ones which are still a little bit broken are the IOS-XR based units as they seem to pass a load of process name, event log number, process number, favourite colour and other random crap in the “message” element. However, they were all a bit broken before all this anyway to be fair. I can see that within the /includes/syslog.php there is a rather extensive section of preg_match/replace for a number of $os types. So I guess the best way forwards to sanitise the extra IOS-XR crap is to build it in there and submit a patch. Although I have a feeling that IOS and IOS-XR count as the same $OS type? So we won't necessarily be able to filter the manipulation using that to match them. Anyway, if we make any progress with it we’ll let you know! Cheers guys, Robert Williams Custodian Data Centre Email: Robert(a)CustodianDC.com http://www.CustodianDC.com

4 5

Graph Netstats
by Darian Jimenez 27 Aug '14

27 Aug '14

Hello, Anyone know where I could get an understanding of how to interpret the graphs for network devices? Is there any documentation to help explain what to look for in devices for packet loss or problems with service? I don't see a way to tell packet drops like you would see on a device normally.

3 3

Re: [Observium] Weird database issue?
by Nick Pais 27 Aug '14

27 Aug '14

Hogan, I'm talking about the current throughput, screenshots attached. Please excuse my "fantastic" GIMPing skills :D Thanks! > Message: 2 > Date: Fri, 8 Aug 2014 17:40:03 +0000 > From: Hogan Whittall <whittalh(a)yahoo-inc.com> > To: Observium Network Observation System <observium(a)observium.org> > Subject: Re: [Observium] Weird database issue? > Message-ID: <6CDB9AE7-7909-401B-80E3-D404CDA371AC(a)yahoo-inc.com> > Content-Type: text/plain; charset="windows-1252" > > If you?re referring to the link speed (100Mb, 1Gb, etc) then that has historically been a problem with permissions when snmpd tries to access ethtool info. You basically have 3 options: > > - Run snmpd as ?root? (Not recommended) > - Enable ETHTOOL_GSET in the kernel > - Add the link speed of each NIC to snmpd.conf (interface ethX 6 <speed in bits/s>) > > I went with the third option and wrote a script to do this for me. This info is kept in the ?ports? table. > > Or are you talking about current throughput and not port speed? > > -Hogan

4 11

fixing graphing of FortiGate memory (used to work, now broken)
by Nathan Phelan 26 Aug '14

26 Aug '14

Hi Guys, Short version: Fortinet memory stops showing in Observium after updating due to Fortinet MIBS not being where Observium expects them. Hack Solution: mv /opt/observium/mibs/FORTINET-* /opt/observium/mibs/fortinet/ ./discovery.php -h all Long version: We have been running a very old version for a long time ( 0.14.1.4945 ) and it's been ticking along nicely but I've finally found some time to do a test update to the latest svn and ran into an issue with Fortigate memory. I ran a discovery with debug; attached as "memory debug.txt". If I run one of those commands manually I get the following: root@clone:/opt/observium# /usr/bin/snmpget -v2c -c 'test -snmp' -OvQ -m FORTINET-FORTIGATE-MIB -M /opt/observium/mibs/rfc:/opt/observium/mibs/net-snmp:/opt/observium/mibs/fortinet 'udp':'testdev.test.local':'161' fgSysMemUsage.0 MIB search path: /opt/observium/mibs/rfc:/opt/observium/mibs/net-snmp:/opt/observium/mibs/fortinet Cannot find module (FORTINET-FORTIGATE-MIB): At line 0 in (none) fgSysMemUsage.0: Unknown Object Identifier (Sub-id not found: (top) -> fgSysMemUsage) The FORTINET-FORTIGATE-MIB files don't appear to have changed but Observium appears to be looking for them under "/opt/observium/mibs/fortinet/" when they are actually just under "/opt/observium/mibs/" After doing this: root@clone:/opt/observium# cd mibs root@clone:/opt/observium/mibs# mkdir fortinet root@clone:/opt/observium/mibs# mv FORTINET-* ./fortinet/ and then running discovery again memory stats came back. Has anyone else had this issue? If so maybe the MIBS need to be relocated in SVN? Cheers, Nathan

2 1

Quote for feature
by Benoit Lambert 25 Aug '14

25 Aug '14

Where can I get a quote for a new feature? The Professional Services (http://www.observium.org/services/) page doesn't contain any information on who to contact. Thanks, Benoit

6 8

Pagination on edit settings?
by Michael Loftis 25 Aug '14

25 Aug '14

Is there *any* way to get pagination added to the edit setting pages? F/ex the "Sensors" page is unusable on a majority of my devices try waiting for load/render when you've got something like a fully loaded Juniper EX9214 so you can disable a sensor alert on one port ... ~300 ports - it's just pure pain assuming it finishes at all. And then you have to set PHP's limits so high to get the submit/save to work. Another issue is I kind of need an ability to occasionally do a "reset all" on the limits page. Individually checking off hundreds of sensors is a completely unnecessary waste of time. I realise I could go muck in the database for both of these in my particular but thats just not always an option and this just seems like something it should be doing when it hits a couple hundred records or some configurable limit. -- "Genius might be described as a supreme capacity for getting its possessors into trouble of all kinds." -- Samuel Butler

2 1

Re: [Observium] Moving logs and rrds to another partition?
by Benoit Lambert 25 Aug '14

25 Aug '14

If anyone is interested, this is how I did it without shutting down the server. And it worked. #Add 2 30GB disk (sdb1=DB | sdc1=rrd) #Rescan to find HDD - without shutdown echo "- - -" > /sys/class/scsi_host/host2/scan #check if it got added fdisk -l ls -la /dev/sd* #Stop daemons service crond stop service snmpd stop service syslog-ng stop service httpd stop service auditd stop #Stop sql last service mysqld stop #Configure DB HDD fdisk /dev/sdb n p 1 <cr> <cr> w #Configure rrd HDD fdisk /dev/sdc n p 1 <cr> <cr> w #Format them in ext4 mkfs.ext4 /dev/sdb1 mkfs.ext4 /dev/sdc1 #Rename DB folder cd /var/lib/ mv mysql mysql_tmp #Recreate mysql folder mkdir mysql #Mount the DB hard drive mount -t ext4 /dev/sdb1 /var/lib/mysql #Copy files over and preserve owners/perms cp -rp mysql_tmp/* mysql/ #change owner/perms of mysql folder chown mysql:mysql mysql #Rename rrd folder cd /opt/observium mv rrd rrd_tmp #Recreate rrd folder mkdir rrd #Mount the rrd hard drive mount -t ext4 /dev/sdc1 /opt/observium/rrd #Copy files over and preserve owners/perms cp -rp rrd_tmp/* rrd/ #change owner/perms of mysql folder chown apache:apache rrd #mount at startup - edit /etc/fstab /dev/sdb1 /var/lib/mysql ext4 defaults 0 0 /dev/sdc1 /opt/observium/rrd ext4 defaults 0 0 #Start daemons service mysqld start service httpd start service syslog-ng start service snmpd start service crond start service auditd start

2 1

Re: [Observium] help with observium
by Benoit Lambert 25 Aug '14

25 Aug '14

Add this to your config.php #PING Settings - Retries/Timeouts $config['ping']['retries'] = 10; // How many times to retry ping (1 - 10) $config['ping']['timeout'] = 1000; // Timeout in milliseconds (50 - 2000)

3 2

help with observium
by Magnus Svensson 25 Aug '14

25 Aug '14

Hi! I got a problem with my new install. i get messege: 2014-08-25 11:00:01 "host" System Device status changed to Up 2014-08-25 10:55:07 "host" System Device status changed to Down (ping) its always down 5min, and if i ping the host its up. its just observium that think its down. ive google it and tried ping settings in config and i dont get it to work. someone know whats wrong? //magnus

3 3

Status Proxmox-VE-Monitoring
by Cyrill Matenaers 23 Aug '14

23 Aug '14

Hi all, I just read something about Proxmox-Monitoring here http://jira.observium.org/browse/OBSERVIUM-471. What is the actual status of this development? Is there any documetation on how to use it? Thanks for your help! Kind regards, Cyrill

4 6