- observium - lists.observium.org

renamed block devices
by Tony Lill 28 Apr '14

28 Apr '14

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Observium CE 0.13.10.4585 doesn't seem to handle renamed block devices very gracefully. I added some disks to one of my boxes, which caused the old sda and sdb to come up as sdb and sdd. Now on the disk i/o page some devices are missing, some appear twice, most are displaying incorrect info. I've attached the contents of the ucd_diskio table. sdc and it's partitions are missing and a number of the others are duplicated. How do I clear up this mess? Ideally I'd like to keep the old data so I could see what effect the changes I'm making has. - -- Tony Lill, OCT, ajlill(a)AJLC.Waterloo.ON.CA President, A. J. Lill Consultants (519) 650 0660 539 Grand Valley Dr., Cambridge, Ont. N3H 2S2 (519) 241 2461 - --------------- http://www.ajlc.waterloo.on.ca/ ---------------- -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iEYEARECAAYFAlMfrVAACgkQGS8yZq1uvxDMvwCff66JvTYJMebz+J6sdq7M5DSy 1DcAnRNL7haZlrK7KF0psYodjVCcmQEA =bMYY -----END PGP SIGNATURE-----

3 13

Device Traffic Overview Graph Legend
by Jake Turner 26 Apr '14

26 Apr '14

Hi all, Just in case anybody else has spotted this and wants to fix it. The problem I was seeing (in the community edition at least) was that when ‘Show Legend’ was clicked on the Device Bits Graph page, the interface names would only have 10 characters as the image below shows: [cid:4474A50E-FFD7-439C-B557-8AB5663E872B@turnerhouse.local] With Cisco devices this renders the legend pretty useless as almost all the interfaces have FastEthernet/GigabitEthernet/TenGig etc. in front of the interface number. To fix this the file /html/includes/graphs/generic_multi_separated.inc.php needs to be modified. Lines 28 & 29 just need more space: $rrd_options .= " COMMENT:'$units_descr Cur Avg Max'"; if (!$nototal) { $rrd_options .= " COMMENT:' Total'"; } Line 81 will have the number 10 after the comma on rrdtool_escape: ($rrd['descr'], 10). This basically strips the interface to that length. In order for the interfaces on my 6500 to show nicely I changed this number to be 25: rrdtool_escape($rrd['descr'], 25). If you decide to go for a different number you should modify lines 28 & 29 too to make sure the ‘Cur, Avg, Max and Total’ headers are in line with the stats below. Then you just need to change for the ‘outbits' as well: substr(str_pad('', 25),0,25). The graph will/should then display like the below. Much more useful :) [cid:937DFBEE-EE5A-4F3E-8B38-AE6466900BE5@turnerhouse.local] An alternative would be to use ifName rather than ifDescr but I can’t be bothered to figure out how to do that. Jake Turner

3 5

Re: [Observium] Alerts
by Adam Armstrong 26 Apr '14

26 Apr '14

1 0

Alerts
by Kenneth Karlsson 26 Apr '14

26 Apr '14

Hi All I have problems configurering alerts, there's no alert mail shooting at me ?! I have this config, I believe I followed instructions, so any help is appreciated: The 10.42.20.49 is a postfix host, and I track in maillog to see any mails/connection, but nothing even though system indicates "Alerted" .. [cid:image001.png@01CF60E5.A6B96960] Med Venlig Hilsen / Best regards Kenneth Karlsson

1 0

Disk I/O feature gone?
by Joshua Elson 25 Apr '14

25 Apr '14

Was this feature removed in the latest community release? We used to be able to view each disk (sda/sdc/sdc) and the ops/sec and bytes/sec. I can't seem to find this anymore after the latest upgrade. Thanks in advance. Josh Elson | Linux Systems Administrator MaxLinear | www.maxlinear.com<http://www.maxlinear.com/> 760-517-1431 | cell # 760-212-6852 2051 Palomar Airport Road, Suite 100 | Carlsbad, CA 92011 [Description: Description: http://wiki/download/attachments/1704274/Maxlinear+Logo.png?version=1&modif…] ________________________________ This electronic mail (including any attachments) may contain information that is privileged, confidential, and/or otherwise protected from disclosure to anyone other than its intended recipient(s). Any dissemination or use of this electronic mail or its contents (including any attachments) by persons other than the intended recipient(s) is strictly prohibited. If you have received this message in error, please notify us immediately by reply e-mail so that we may correct our internal records. Please then delete the original message (including any attachments) in its entirety.

5 8

Dell (Eaton) UPSs
by Pedrosi, Derek G. 25 Apr '14

25 Apr '14

Any chance for Dell UPSs (rebranded Eatons) being added? MIBs attached. sysObjectID -> .1.3.6.1.4.1.674.10902.2 derek in Detroit _______________________________________________________________ NOTICE TO PERSONS SUBJECT TO UNITED STATES TAXATION (MCPS) DISCLOSURE UNDER TREASURY CIRCULAR 230: The United States Federal tax advice, if any, contained in this document and its attachments may not be used or referred to in the promoting, marketing or recommending of any entity, investment plan or arrangement, nor is such advice intended or written to be used, and may not be used, by a taxpayer for the purpose of avoiding Federal tax penalties. _______________________________________________________________

1 0

Anyone ever seen something like this after updating ?
by Tom Taylor 25 Apr '14

25 Apr '14

http://imgur.com/Xu5JMhn Just upgraded and it looks like all the image rendering's gone wonkey. Any ideas ? -TT Tom Taylor Head of Engineering, Chicago T +1 312 605 8900 D +1 312 605 8907 [http://www.themill.com/logos/MillLogo.png] The Mill 54 West Hubbard Street, Suite 502, Chicago, IL, 60654 themill.com | <http://www.themill.com/> themillblog.com | <http://www.themillblog.com/> @millchannel | <https://twitter.com/millchannel> facebook.com/millchannel<http://www.facebook.com/millchannel>

1 1

Failover interface monitoring
by Gregoire Tourres 25 Apr '14

25 Apr '14

Hello Observium team I am monitoring different kind of devices (linux / windows / cisco) through my Observium. Everything work fine except for some Windows server 2012 network interfaces : the failover interfaces. Microsoft uses some kind of interface to supply a failover service (2 servers of the same kind speaks a special black protocol to monitor themself). >From the Windows Server, interfaces are "Microsoft failover cluster virtual adapter". >From Observium, they are seen as ethernet device and as tunnel device. Observium always seens some packet loss on these interfaces (however, there are no packet loss in the network, even with huge stress). So ... it keeps alerting for packet loss. You would say "turn monitoring or alerting off on these interfaces". Well .. that what I did : But I keep getting some alert in the eventlogs : The others real interfaces on the server work fine (good monitoring, no loss, no alerting). I am only having the problem on these kind of interfaces. Why does Observium always see packet loss on these interface ? Why does Observium do not stop alerting when I tell him to ? Thanks for answer! Greg

3 2

VM and Observium
by Luis Quintana 25 Apr '14

25 Apr '14

So, is anyone experiencing any issues running on observium on a vm? I submitted an email a month or two ago about some issues we were seeing, and i decided to reinstall observium from scratch. Currently i am seeing the same issues, whereby it "appears" that connectivity drops or isnt captured randomly. it's not across one device, but all. i am wondering if maybe i didnt build the vm appropriately or we really are having a network type issue. ive looked at the network but maybe i am overlooking something. anything on the vm side that maybe i can look at? i am running ubuntu 13.04 and the latest observium code as of yesterday.

4 13

Brocade MLXe and ICX support / question
by Swen Veckes (KDDI) 25 Apr '14

25 Apr '14

Hi, we use a couple of MLXe’s and ICX’s which seem fine at the first place running observium’s subscriber edition. However, a few part seems not working correctly: + MLXE’s - Memory shown as -1 on the device main page and has no graphs „nan“ at all - Memorypool usage graphs show -nan - Observium recognizes the installed Line card CPU but report 100% CPU on another machine adder 3 month earlier it reports OK - VLAN’s are reported fine, however, on the VLAN tab only untagged vlan1 (default vlan) has port associated with it, others are empt + ICX’s - CPU reported 100% all time, but same switch installed earlier shows fine - Compared to the MLXe it does not show LAG interfaces Any hints on the the graphing which works for one device but another added later does not as well for the memory? I Don’t hink that observium suddenly broke the support for this, as it is still working for the other devices (exactly the same device). What do I need to provide to help fixing the other things such as the LAG display for the ICX or others and to get more current Brocade device support? Cheers, Swen

3 7