Channels

Info

Credits

  • cheeser
  • joed
  • kinabalu
  • lunk

2008-04-30

#opennms 2008-05-01

2008-05-02

Nick Message Date
*** HTI :142.179.92.143 joined the channel [00:58]
*** PowerOp :116.59.12.179 joined the channel [01:37]
*** HTI parted the channel [01:43]
*** parag0n :82-70-62-126.dsl.in-addr.zen.co.uk joined the channel [02:33]
*** parag00n :82-70-62-126.dsl.in-addr.zen.co.uk joined the channel [02:38]
*** exothermc :74.85.89.146 joined the channel [03:52]
*** sartin :pdpc/supporter/active/sartin joined the channel [05:43]
<sartin> morning chaps [05:45]
*** mhuot :pdpc/supporter/active/mhuot joined the channel [06:11]
<sartin> mhuot! [06:17]
<mhuot> sartin! [06:18]
*** _indigo :dslb-088-068-184-155.pools.arcor-ip.net joined the channel [06:26]
*** chalco_lab :about/networking/255.255.255.240/chalco joined the channel [06:42]
*** _indig1 :dslb-088-068-184-155.pools.arcor-ip.net joined the channel [06:45]
*** nmsmanjohn :rrcs-24-106-176-34.se.biz.rr.com joined the channel [07:08]
*** gvainfo :zehill.genevainformation.ch joined the channel [07:20]
*** jeffg :pdpc/supporter/active/jeffg joined the channel [07:21]
<gvainfo> jeffg! [07:32]
<jeffg> gvainfo! [07:32]
<_snd> _gva! [07:35]
<_snd> _jef! [07:35]
*** k-tr :DSL01.83.171.144.168.ip-pool.NEFkom.net joined the channel [07:44]
<McFly> Anyone have any idea what can be done about this? http://paste2.org/p/24520 [07:58]
<_snd> to approaches: file a bug and get on to dj, or mkfs :) [07:59]
<_snd> two, even [08:00]
<RangerRick> yeah, looks like that device is sending some malformed interface information? [08:00]
<RangerRick> open a bug, and include a raw snmpwalk of that device [08:01]
<McFly> yeah, bad (or missing) speed info [08:01]
<McFly> IF-MIB::ifSpeed.3 = Wrong Type (should be Gauge32 or Unsigned32): "" [08:01]
<McFly> I guess that'd do it :/ [08:01]
*** mwmwmw parted the channel [08:01]
<jeffg> McFly, have a 2x4 ready the next time that vendor's SE comes to visit [08:08]
<McFly> They're old units, I suspect there's no fix from that end whether I use 2x4 or not. OpenNMS would have to handle that condition [08:10]
<_snd> the 2x4 condition? [08:10]
<McFly> heh [08:11]
<jeffg> catch twoByFourException e { /*silently discard */ } [08:12]
<_snd> catch twoByFourException e { throw fit } [08:13]
<jeffg> McFly, we could probably code a workaround for an empty ifSpeed. there's no guarantee the device's agent isn't flaky in other ways, of course, but it could get you over the hump [08:13]
<jeffg> _snd, throw new Fit(); [08:13]
<McFly> Well, it's actually working for what I need it to do with OpenNMS, so it's not really a huge deal...just hate the annoying daily exception from capsd on the 9 units I have. [08:14]
<jeffg> McFly, coolio, just so you're not scared to open a ticket :) [08:15]
*** parag000n :82-70-62-126.dsl.in-addr.zen.co.uk joined the channel [08:19]
*** fskrotzki :host198.textwise.com joined the channel [08:22]
*** mwmwmw :198.246.149.100 joined the channel [08:40]
<djgregor> McFly: open up a bug on that; we shouldn't crap ourselves on bad data [08:40]
<McFly> bug 2448 opened [08:41]
<djgregor> McFly++ [08:41]
<sfb> http://www.zabbix.com/screenshots_11_overview.php [08:41]
<sfb> That's a nifty view. [08:41]
*sfb should open an enhancement request. [08:41]
<sfb> haha [08:42]
*_snd wanna have like CA Uniwhatever TNG: 3d surf view! [08:44]
<jeffg> the CA UniCrapper dragon? [08:50]
<_sndbot> nothing like wasting time repeating the same ip range? [08:50]
<jeffg> ~joed++ [08:50]
<_sndbot> joed has a karma level of 16, jeffg [08:50]
*** p-brane :server1.opennms.org joined the channel [08:51]
<_snd> jeffg: somthing like that, whatever p-brane called it [08:51]
<djgregor> McFly: can you apply a patch, recompile, and try it out? [08:52]
<McFly> no [08:53]
<djgregor> okie [08:53]
<McFly> Sorry, I know that doesn't help, but I've no idea how to do that [08:53]
<djgregor> no problem at all [08:53]
<djgregor> it's non-trivial unless you've done it a good number of times [08:54]
<p-brane> shart? [08:58]
<djgregor> p-brane: that's easy... just sneeze when you have a stomach bug [08:59]
<p-brane> lol [09:03]
<p-brane> djgregor: so I can clearly not choose the wine in front of you [09:21]
<_snd> chalkie :) [09:30]
<sfb> Hrm [09:50]
<sfb> I had snmp-graph.properties [09:51]
<sfb> er [09:51]
<sfb> I hate it. [09:51]
<sfb> How can I prevent the netsmp.loadavg graph from converting 0.03 to 300 m [09:52]
<McFly> Remove the CDEF? [09:54]
<jeffg> sfb, remove the "%s" from the GPRINT [09:54]
<jeffg> if you do that you may need to increase the precision (%8.2lf gives only two digits to the right of the decimal point) [09:55]
<sfb> Okay [09:56]
<sfb> So [09:56]
<sfb> UCD-SNMP-MIB::laLoadInt takes your load average (0.03, in this case) and multiplies it by 100 which gives me a 3. [09:57]
<sfb> Then in snmp-graph.properites I have: CDEF:float1=avg1,100,/ [09:57]
<sfb> Which results in 0.03 [09:57]
<sfb> But I see like 300 m in the graph... So it's converting it's unit of measure somewhere. [09:58]
<sfb> So removing the %s means it won't convert the unit of measure? [09:58]
*sfb tries.. [09:59]
*** MrCerulean :209-203-68-2.static.twtelecom.net joined the channel [10:00]
<jeffg> sfb, %s means "SI units" [10:02]
<jeffg> and removing %s from the GPRINT affects only the GPRINT -- i'm not sure how to make the vertical axis units not be SI, though i bet it's possible. [10:02]
<jeffg> MrCerulean! [10:02]
<MrCerulean> jeffg! [10:04]
<sfb> jeffg: That worked great, thanks! [10:05]
<jeffg> sfb, welcome! [10:05]
<sfb> The graph is correct, the legend was incorrect. [10:05]
<p-brane> You rush a miracle man you get rotten miracles. [10:06]
<jeffg> heh [10:07]
<jeffg> you p-brane, ever used embedded flag expressions with java.util.regex.Pattern? [10:24]
<jeffg> (s/you /yo /) [10:24]
<jeffg> ooh, i think i just answered my own question [10:25]
<jeffg> yep [10:25]
*** Sortova :nc-63-162-204-51.sta.embarqhsd.net joined the channel [10:26]
<p-brane> jeffg: I know not what you speak unless I did it without knowing... which I probably have ;-) [10:26]
<jeffg> p-brane, is this article now obsolete since we've got the groovy event-forwarding thing? http://www.opennms.org/index.php/Hyperic_HQ_Syslog_Integration [10:26]
<jeffg> p-brane, e.g. Pattern.compile("(?s)foo.*bar") [10:27]
<sartin> anyone familiar with the Ticket interface? [10:29]
<jeffg> sartin! [10:29]
<sartin> jeffg! [10:30]
*sartin has written the SOAP ticket interface for OTRS now [10:30]
<jeffg> sartin, somewhat familiar, but i'm certainly not the best resource [10:30]
<p-brane> jeffg: yes, obsolete [10:31]
<sartin> I want to add a new value to State, and wonder what the impact might be. [10:31]
<p-brane> sartin: coolio [10:31]
<_snd> iDad :) [10:31]
<p-brane> sartin: hmm [10:31]
<jeffg> p-brane, obsolete? [10:31]
<p-brane> sartin: I started down that path and I have a recommendation [10:31]
<p-brane> jeffg: the link [10:31]
<p-brane> sartin: that state needs to map to the NGOSS TT state standard list [10:32]
<jeffg> p-brane, oh, sorry, that fell out of my state table :) thanks [10:34]
<MrCerulean> http://www.youtube.com/watch?v=05Nj72C82io [10:40]
<p-brane> sartin: http://www.opennms.org/documentation/XmlTroubleTicketSchema.xsd [10:44]
<jeffg> RangerRick, do you know of a way to mark an article as obsolete in mediawiki? [10:44]
<RangerRick> we should probably make a template if there isn't one already [10:44]
*RangerRick makes one [10:45]
<sartin> p-brane: thanks. [10:48]
*** Markw78 :216-54-141-250.static.twtelecom.net joined the channel [10:48]
<sartin> p-brane: gulp! [10:50]
<p-brane> sartin: there at 2 states that map the 2 states we use in opennms [10:52]
<p-brane> "RequestState" and "TroubleState" [10:52]
<RangerRick> jeffg: ok, I added one... see http://www.opennms.org/index.php/Sandbox for an example [10:52]
<p-brane> pretty short list [10:52]
<jeffg> RangerRick, thanks! [10:52]
<p-brane> sartin: there is the troubling status list, though: "TroubleStatus" (yikes) [10:53]
<RangerRick> jeffg: and the 2nd argument is optional [10:55]
<jeffg> RangerRick, sweet! [10:55]
<jeffg> ~RangerRick++ [10:55]
<_sndbot> rangerrick has a karma level of 30, jeffg [10:55]
*RangerRick (hearts) mediawiki [10:56]
*** bandini :host22-23-dynamic.20-79-r.retail.telecomitalia.it joined the channel [11:00]
<jeffg> huh, who knew? http://upsidedowntomatoplant.com/ [11:01]
<p-brane> jeffg: I checked those out earlier this week and thought it would be cool to build one. [11:10]
<jeffg> RangerRick, good catch on 2434, thanks [11:15]
<RangerRick> I was gonna merge it, since it's one of the few that I didn't commit, but then I saw it depended on a change I did commit [11:15]
<RangerRick> so I just updated the bug instead ;) [11:15]
<RangerRick> probably better if someone that understands the thresh code a bit more commits it anyways [11:16]
*RangerRick just sent an email to ogp [11:16]
<RangerRick> which also happens to show how you query unverified stuff using the new milestones [11:16]
<RangerRick> I'm gonna work on the release notes [11:17]
<p-brane> oh, cool: http://www.springsource.com/web/guest/products/suite/applicationplatform [11:19]
<RangerRick> ooh, osgi framework [11:21]
<p-brane> Tomcat [11:27]
<p-brane> ;-) [11:27]
<jeffg> RangerRick, "mayday" you think? [11:36]
<RangerRick> sure [11:36]
<MrCerulean> jeffg: Philosophical question? [11:44]
<jeffg> MrCerulean, it could be, but i was suggesting a codename for the 1.5.91 release :) [11:45]
*jeffg pense, donc il est [11:45]
<MrCerulean> jeffg: No. I *have* a philosophical question if you have a minute. [11:46]
<MrCerulean> jeffg: :) [11:46]
<jeffg> MrCerulean, philosophize at will [11:54]
<jeffg> you lover of wisdom you [11:54]
<_sndbot> in the table. We just get adopted by the people moved to hibernate [11:54]
<MrCerulean> jeffg: Say I want to look at high load average thresholds. However, I only want *notifications* on certain groups of systems. Should I manage that by refining the *threshold* or by configuring different *notifications*? [11:57]
<jeffg> MrCerulean, i would tweak the notifications, given what you've told me so far [11:58]
<MrCerulean> jeffg: OK. Given that, then, if I want to notify based on Surveillance Group membership, what's the syntax? [11:59]
<jeffg> MrCerulean, to check for membership in a single surveillance category, use categoryName = 'foo' [12:00]
<jeffg> to check for the intersection of surveillance categories foo and bar, use: catincfoo & catincbar [12:00]
<MrCerulean> jeffg: I have this: "(IPADDR != '0.0.0.0') & (categoryName=='Login') & (categoryName=='P&R') & (categoryName=='RFIC')" [12:01]
<MrCerulean> jeffg: And it doesn't seem to work the way you'd think. [12:01]
<Markw78> I am looking for a specific response code from an HTTP Get and need to track how often it occurs...Any ideas? [12:03]
<nmsmanjohn> you could write a script to look for the response, and output a "1" or a "0" for the condition. [12:04]
<nmsmanjohn> then you could graph it. [12:04]
<jeffg> MrCerulean, you need to use catinc [12:04]
<jeffg> MrCerulean, also the ampersand in 'P&R' will be problematic [12:05]
<jeffg> you may need to replace it with &amp; [12:05]
<jeffg> and it may throw a wrench into the whole works -- personally i would rename that category to "PandR", using an ampersand in a label like that is just asking for mysterious trouble down the road [12:06]
*** brozow :cpe-076-182-023-094.nc.res.rr.com joined the channel [12:07]
<jeffg> MrCerulean, if you change that category name (which you'd need to do via SQL, UPDATE categories SET category_name = 'PandR' WHERE category_id = ?) then be sure that you update surveillance-views.xml if it's used in there [12:08]
<nmsmanjohn> Question: Someone here sent me a link that showed some Groovy scripts used with ONMS. I messed up and didnt book it. I'm looking for a script that created an event in ONMS from Groovy. [12:08]
<nmsmanjohn> i've been searching thru the project but cant find it (search foo not working too well). [12:09]
<nmsmanjohn> brozow: afternoon. [12:09]
<brozow> how you doing nmsmanjohn ? [12:10]
<nmsmanjohn> heh [12:11]
<nmsmanjohn> ok. [12:11]
<nmsmanjohn> working om my Groovy script fo rONMS. [12:11]
<p-brane> http://www.opennms.org/index.php/Groovy_Scripting [12:11]
<brozow> nice [12:11]
<MrCerulean> jeffg: OK. I changed it, and (IPADDR != '0.0.0.0') & (catincPandR) gives you what you'd expect. [12:12]
<nmsmanjohn> thanks p-brane but I have that one. The one I'm looking for was actualy "IN" in sourceforce project. [12:12]
<nmsmanjohn> i got it o get the ips. poll them and can look at the results. [12:13]
<nmsmanjohn> I next have to add a category, and create an event when something is wrong. [12:13]
<MrCerulean> jeffg: Also, (IPADDR != '0.0.0.0') & (catincLogin) gives you what you'd expect. [12:13]
<nmsmanjohn> brozow: where is that Groovy script that you did that would create an alarm/event? [12:14]
<MrCerulean> jeffg: However, (IPADDR != '0.0.0.0') & (catincPandR) & (catincLogin) probably needs an 'or' instead of an 'and'. [12:14]
<brozow> there are a few scripts in opennms-tools/groovy-somethingorother/src/main/groovy [12:15]
<mwmwmw> liferay is a pain in the ass, as is opennnms to get AD auth to work. [12:15]
<nmsmanjohn> thats where it is. [12:15]
<jeffg> MrCerulean, that could well be -- if those two categories don't intersect, it's definitely what you need to do. [12:15]
<brozow> mwmwmw: ever think that maybe the problem is AD? [12:16]
<mwmwmw> brozow, maybe. :) [12:16]
<MrCerulean> jeffg: Yeah. I'm trying to say I only want to see high load average alerts for machines in the following groups... [12:16]
<mwmwmw> brozow, but hte AD guys claims everything is rather standard. [12:16]
<jeffg> mwmwmw, have him check the source of the AD server to be sure [12:16]
<sfb> hahahah [12:17]
<mwmwmw> jeffg, you know very well, most business just blindly use AD. [12:17]
<sfb> AD isn't "rather standard." [12:17]
<nmsmanjohn> Thanks borzow. [12:17]
<mwmwmw> as far as AD is for AD, it's rather standard "AD". [12:17]
<sfb> Microsoft in typical fashion uses their own custom schema and have their own quirks on their LDAP. [12:17]
<nmsmanjohn> p-brane: it was http://opennms.svn.sourceforge.net/viewvc/opennms/opennms/trunk/opennms-tools/groovy-tools/src/main/groovy/ [12:17]
<sfb> And their Kerberos is damn near totally incompatible with standard Kerberos servers. [12:17]
<sfb> mwmwmw: Gotcha. [12:17]
<sfb> I want to set up RedHat Directory Services (Fedora DS, actually, but still...) [12:18]
<sfb> Just to tinker with. [12:18]
<mwmwmw> also, the lack of docs even for liferay, is rather sad. [12:18]
<mwmwmw> I just deployed liferay 5, to see if by some chance in hell they worked around some bugs. [12:18]
<sfb> What is Liferay? [12:23]
<mwmwmw> holly poop!. same settings, liferay 4.4.2..... put the same crap in 5.0.1rc.... worked. [12:23]
<sfb> I mean "portal" is kind of ambiguous. [12:23]
<mwmwmw> HA! [12:23]
<MrCerulean> jeffg: OK. Got all that working properly. One last question: Now that I have my High Threshold notifications set to only fire when a machine belongs to a certain group, how do I deal with the fact that I need to check memory on a completely different group? Create a new threshold? [12:24]
<mwmwmw> sfb. I've never messed with a portal before.... :) never needed to. but it's supposed to replace our html 1.0 compliant circa 1999 intranet stuff with bash and perl scripts glueing everythign to gether. [12:24]
<jeffg> MrCerulean, what are your existing high thresholds checking? [12:26]
<MrCerulean> jeffg: loadavg5, lmTempSensorsValue, cpuBusyTimePct [12:27]
<jeffg> MrCerulean, well, memory is a different resource, so yeah, i think you need to define a new threshold. unless i'm misunderstanding you somewhere, which i could be. [12:28]
<MrCerulean> jeffg: I guess the meta question is: What's the best practice for checking different thresholds across different groups of machines? [12:29]
<jeffg> MrCerulean, depends totally on how you group those machines :) [12:30]
<brozow> RangerRick: know an easy way to send a growl message from the command line? [12:36]
<brozow> or anyother message that will pop up and interupt me? [12:36]
<RangerRick> brozow: I think you could do it with osascript, but I don't know the syntax [12:36]
<brozow> I'm just looking for a smple command I can append to the end of a long comnpile that will say 'hey the compile is done' [12:37]
<RangerRick> brozow: http://www.macosxhints.com/dlfiles/growl_sh.txt [12:41]
<brozow> sweet! you rock RangerRick [12:42]
<RangerRick> hah, gotta dos2unix it [12:42]
<RangerRick> woot! [12:43]
<RangerRick> it works [12:43]
<brozow> is perfext [12:44]
<brozow> I needed thanks [12:44]
<brozow> ^just [12:44]
<brozow> RangerRick: what happened to 1.6.. doesn't seem to have merge info for 1.6-testing anymore [12:49]
<brozow> it seems like it only just happened in the last half hour [12:49]
<RangerRick> brozow: erm? lemme look [12:49]
<_snd> its dj's fault [12:54]
<jeffg> duh [12:54]
<_sndbot> opennms enterprise edition we added a new Date() (now) and one for your phone: http://stopabductions.com/ [12:54]
<RangerRick> brozow: I did an svn up and svnmerge-integrated still has it [12:55]
<RangerRick> /opennms/branches/1.6-testing:1-8497,8499-8556,8558-8765,8767-8787,8789-8959,8964,9003-9004,9006,9010,9022,9025-9027,9029 [12:55]
<RangerRick> brozow: maybe a local merge broke it? a conflict or something? [12:56]
<brozow> let me check [12:58]
<brozow> yeah that was it.. I had mid air collision with your last merge... [12:59]
<brozow> but I had reverted and somehow '.' get missed I guess. [12:59]
<brozow> sorry for that [12:59]
<RangerRick> I tend to do "svn revert -R ." to start over [13:02]
<brozow> hmm... that is what I did... must have dorked something else [13:05]
*brozow seems to remember hitting Ctrl-C for something related to this [13:05]
<fastjay> morning [13:06]
<adeel> morning fastjay [13:06]
<geoaxis> screen -x [13:10]
<geoaxis> ouch .... [13:10]
<geoaxis> i am already in irssi [13:10]
<adeel> hahah [13:10]
<brozow> fastjay! [13:10]
<brozow> fastjay: just yesterday I got a book called 'RESTful Web Services' [13:11]
<chiadog> mmm, nouns [13:11]
<geoaxis> brozow: Orieally? [13:12]
<geoaxis> O'rielly* [13:13]
<chiadog> o rly? [13:13]
<fastjay> brozow: nice [13:14]
<geoaxis> i am referring to a publisher [13:14]
<fastjay> i am thinking restd should really be part of the webapp.. but.. also realizing i have no time to make that a realization. Best practicies say your urls should be able to be cut from your browser to do the thingy [13:14]
<fastjay> we keep running into issues w/ zenoss (more so the person implimenting than the app) [13:15]
<fastjay> i soo wnat to put opennms in :) [13:15]
*** Dalek3 :66-202-165-66.rev.knet.ca joined the channel [13:21]
<MrCerulean> jeffg: They're currently grouped into Surveillance Categories. [13:21]
<Dalek3> hi all [13:21]
<MrCerulean> jeffg: But I'm happy to group them in whatever fashion they need to be grouped in. [13:22]
<Dalek3> I have been playing with opennms and I have a question [13:22]
<jeffg> i was just cleaning it, and it went off! [13:22]
<_sndbot> that was the way, our tests stopped cleaning up [13:22]
<Dalek3> can it handle managing two different computers that have the same IP address without lumping them together? [13:23]
<jeffg> Dalek3, yes, duplicate ip addresses do not make opennms' head explode, but it will obviously have difficulty polling services on both nodes. [13:24]
<Dalek3> I'm using the PSK [13:24]
<jeffg> oh wow, you've been doing some serious playing then [13:25]
<Dalek3> these are for multiple networks behind their own nat routers.. the systems on them send me traps from the anti virus software [13:25]
<Dalek3> I had to configure my own event xml for the anti-virus software because the vendor's didn't seem to work with mib2opennms [13:25]
<jeffg> Dalek3, did you run smilint to try to isolate the problem with the vendor mib? [13:26]
<Dalek3> no, but I downloaded a few other mibs and they all did the same thing [13:30]
<Dalek3> I have a feeling it's because of the ancient mib2opennms version that came with it [13:30]
<Dalek3> but there are only 4 traps so it wasn't a big deal to make an event.xml manually [13:30]
<Dalek3> using the vmware version which runs in mandriva, the only mandriva package for mib2opennms was version 0.3.0 [13:31]
<jeffg> heh, 0.3.1 is the latest i think [13:31]
<Dalek3> yeah [13:31]
<jeffg> it's "tricky" to get working right [13:31]
<Dalek3> so anyhow I'm trying to get those sophos traps to trigger creating the node and interface [13:31]
<Dalek3> but I can't tell where the output of the autoaction command is going.. I see "rc = 2" in the action log and of course a non zero return code is generally bad [13:33]
<Dalek3> so I can't tell what the error is because I can't figure out which log it outputs it to if any [13:34]
<RangerRick> 0.3.1 was only a build-system change and one int cast [13:35]
<RangerRick> so it's unlikely to have any effect on your issues :) [13:35]
<Dalek3> plus I manually created one host and interface using the send-event.pl [13:36]
<Dalek3> now when I try creating another host with an IP I previously used it doesn't seem to work [13:37]
<Dalek3> I tried deleting the host but it doesn't actually seem to get deleted, more like just flagged as deleted, and snmp traps still get associated with it even though it's been deleted [13:38]
<Dalek3> which I find a bit confusing [13:38]
<Dalek3> unless I have to delete the interface too? [13:39]
<jeffg> Dalek3, yes, when you delete a node it just gets marked as such, and an automation runs daily that does the actual deletion. cascading deletes in the db schema take care of the interfaces and other associated objects. [13:41]
<jeffg> a better way to create nodes that you can't reach directly is probably to use provisioning groupsw [13:42]
<jeffg> groups even [13:42]
<Dalek3> I saw the provisioning groups but coudln't tell what they did [13:44]
<Dalek3> and there didn't seem to be any help on that on the site [13:45]
<jeffg> it's a front-end to the model importer, which allows you to "outsource" some or all of opennms' notion of the nodes it manages. you can use it with or without capsd. [13:46]
<Dalek3> right but wouldn't I have to create the nodes manually then? [13:49]
<sartin> Eldest' birthday today. 6 years old. [13:55]
*sartin is exhausted [13:55]
*djgregor sends sartin a nightcap [14:00]
*sartin thanks djgregor [14:02]
<sartin> 'nother question on integration with OTRS. Add on module (as other ticketerplugins) or built as standard? [14:04]
<sartin> OTRS is all GPL as far as I can see? [14:04]
*McFly sure would be interested in seeing that integration when done. We're using OTRS now and I've 'integrated' it when a few standalone scripts and alarm/detail.jsp hacking [14:06]
*Markw78 slaps mwmwmw around a bit with a large trout [14:08]
<Markw78> I got my AD auth working if you have any Q's [14:08]
*** mhuot :pdpc/supporter/active/mhuot joined the channel [14:10]
<sartin> McFly: won't make it in for the next release, but will be available for testing first half of the month. [14:18]
*** `nstuff :c3p0.reverse.net joined the channel [14:21]
<`nstuff> evening [14:21]
<`nstuff> how can i set snmp from 'not monitored' to monitored ? [14:22]
<p-brane> `nstuff: poller-configuration [14:24]
*** mpeterse1 :66.88.39.226.ptr.us.xo.net joined the channel [14:28]
<archon-> http://en.wikipedia.org/w/index.php?title=Comparison_of_file_systems&oldid=209063556#Features [14:28]
<Dalek3> http://www.opennms.org/index.php/Passive_Status_Keeper [14:30]
<Dalek3> I find that document quite confusing.. as in what comes first, the node or the service [14:30]
<Markw78> So right now I get a lot of alerts for my web sites from NMS poller... Any way I can make it tell me why it timed out, or if it got an error code back from HTTP? I suspect they are HTTP 502's but all ONMS tells me is "down" [14:34]
<Markw78> thoughts suggestions? I'm not much of a coder/scripter =x [14:34]
<p-brane> Dalek3: you have to define the service in capsd-configuration first. [14:34]
<p-brane> Dalek3: then you can have it automatically added to a node during a capsd scan or you can provision a node with that service in provisioning groups. [14:35]
<p-brane> You have to restart after you create the service in capsd-configuration [14:35]
<Dalek3> ok cool, it looks like I will need to do it the same type of way as the windows services one [14:38]
<`nstuff> how would i add lmsensor to smnp ? [14:38]
<jeffg> hmm... cisco mars may win the award for most embarrassingly named process: "pnesloader" http://ciscomars.blogspot.com/2006/10/unable-to-send-request-for-processing.html [14:40]
<p-brane> `nstuff: not sure what you are asking [14:40]
<p-brane> jeffg: LOL [14:41]
<`nstuff> got lmsensor running on afew boxes, how would i add it to opennms [14:41]
<mpeterse1> Did maven stop auto-downloading modules with 1.5.90? [14:44]
<mpeterse1> or artifacts... [14:44]
<p-brane> `nstuff: I'm not familiar with lmsensors by I googled this: http://www.netstatz.com/xenonms/Debian_ONMS_HOWTO_2.4-13.html [14:45]
<Dalek3> p-brane OK I just edited the capsd-configuration.xml, added the service and I'm restarting [14:49]
*** bioanarchism :unaffiliated/bioanarchism joined the channel [14:50]
<mpeterse1> nm, it seems to be downloading now. maybe I had an internet glitch or something. [14:58]
<p-brane> mpeterse1: cool [14:59]
*** moj1 parted the channel [15:00]
<mpetersen> p-brane: as long as everything is on different ports, and reason I can't run 2 opennms instances pointing to different back-end servers with jetty? [15:03]
<mpetersen> so jettyB -> onms-server1 & pgsql db opennms and jettyB -> onms-server2 & pgsql db opennms-dev. [15:04]
<mpetersen> same db/gui server, different onms "main" servers, or whatever you call the server opennms runs on... [15:05]
<mpetersen> I had troubles trying to do this with tomcat in the past... [15:05]
<mpetersen> that should be jettyA and jettyB or jetty and jetty-dev, not 2 jettyB's ;) [15:05]
<p-brane> mpetersen: I'm not smelling what you are stepping in ;-) [15:05]
<mpetersen> I see... [15:06]
<mpetersen> let me try that again. [15:06]
<mpetersen> currently I have pgsql on server1, opennms on server2 and jetty on server3. I want to have pgsql run two databases on server1 (opennms and opennms-dev), 2 opennms instances (server2 and server2-dev) and have jetty and jetty-dev on server3. jetty:8980 and jetty-dev:8981 or similar... [15:07]
<mpetersen> so I can test some config changes and stuff with out messing up the original instance... [15:07]
<p-brane> mpetersen: at RS we have 3 opennms servers configured connecting to one Jetty instance and 1 DB. [15:07]
<mpetersen> I couldn't do this with tomcat, ran into issues... [15:07]
<mpetersen> I don't necessarily want to clutter up the gui by mergeing the 2 opennms servers though... [15:08]
<mpetersen> is that what happens at RS? [15:08]
<p-brane> capsd and poller on 1, collector on another [15:08]
<mpetersen> oh... [15:09]
<mpetersen> I just want a dev instance, but I want to share the gui box (mostly idle) and the DB box (not as idle, but not super busy...) [15:09]
<mpetersen> the rest I'm sure I can handle... [15:09]
<mpetersen> I couldn't do this with Tomcat though.. Tarus tried to get it working and said I'd need two tomcat instances which sounded like a pain... [15:10]
<chiadog> huh, seems odd you couldn't just spin up two webapps [15:11]
<chiadog> wonder what the conflict was [15:11]
<mpetersen> Some global parameter I think... [15:12]
<mpetersen> I figued with jetty I'd have two instances, but less hastle than 2 tomcat instances so it would probably work... [15:13]
<chiadog> not too hard to spin up multiplate tomcats...waste of resources tho [15:13]
<p-brane> mpetersen: I don't know for sure exactly what you want to do but if you are sharing a DB you could have conflicts with 2 different servers monitoring the same nodes... it is possible a couple of different ways but it's not flip the switch kind of thing... not sure what the Tomcat issue was. [15:13]
<p-brane> If there was a webapp challange, I don't think it would matter if it was Jetty or Tomcat. [15:13]
<chiadog> right [15:13]
<chiadog> or would be surprised if jetty somehow isolated contexts better than tomcat [15:14]
<mpetersen> Well if I have 2 jetty's going, wouldn't it be similar to 2 tomcats? I actually want to monitor the same nodes twice ;) [15:14]
<mpetersen> whereas I couldn't get two instances of the same tomcat running two different versions of OpenNMS. [15:15]
<mpetersen> but only the dev servers... I want to try new versions of OpenNMS in a similar distributed setup without needing 3 new servers... [15:15]
<mpetersen> I have 1 spare server I can run the core opennms stuff on. [15:15]
<mpetersen> I could do it all with Xen on the one server, might be close enough to the same I guess... [15:16]
<Sortova> wow - CSS can be pretty powerful: http://nedbatchelder.com/blog/200805/css_homer_animated.html [15:16]
<mpetersen> Sortova: javascript with CSS... it didn't work for me at first, had to enable java script... [15:18]
<mpetersen> Sortova: so do you remember what my tomcat issue was with 2 instances? [15:19]
<chiadog> wow, what an immense waste of time [15:19]
<RangerRick> awesome and scary at the same time [15:19]
<Sortova> mpetersen, not sure [15:20]
<Dalek3> ok I created the service, sent a trap from a system, but it doesn't appear to be creating the node for it or anything [15:20]
<Sortova> mpetersen, can you remind me of the circumstances? I'm still a bit under the weather and not firing on all cylinders today [15:20]
<Markw78> I learned SNMPGET with PHP... it would be cool if ONMS had something like that built in... a table you can customize and populate for viewable SNMP data [15:21]
<mpetersen> Oh.. I wanted to have a onms-gui:8180/opennms and a onms-gui:8180/opennms-dev and for some reason I'd need two tomcat installs to get it working... [15:21]
<p-brane> mpetersen: ah, that makes more sense [15:22]
<mpetersen> but since I changed to jetty, I want to get onms-gui:8980/opennms and onms-dev-gui:8981/opennms running on the same server.. I think it should work with jetty, just set the ports properly? [15:22]
<p-brane> mpetersen: yeah, we've tried this before and couldn't get it done with tomcat [15:22]
<mpetersen> and run two instances of jetty... [15:22]
<p-brane> mpetersen: I don't know, even with Tomcat it seemed feasible [15:22]
<mpetersen> p-brane: ahh.. I was jiving it wrong... I tend to do that around here ;) [15:23]
<p-brane> I hacked on that for a day at Swisscom and never got it working [15:23]
<p-brane> It's worth a shot again [15:23]
<mpetersen> you hacked on single tomcat, 2 instances? or jetty 2 instances? [15:23]
<Dalek3> and the service isn't getting associated with the traps either [15:24]
<p-brane> mpetersen: Tomcat about quite a while back after we changed the RRD default RRAs [15:24]
<p-brane> between 1 and 2 years ago [15:24]
<Dalek3> this thing is driving me nuts [15:24]
<gvainfo> jeffg [15:25]
<jeffg> gvainfo? [15:25]
<gvainfo> jeffg I am laughing my *** off [15:25]
<gvainfo> and I'm sitting with my wife + in laws [15:25]
<jeffg> laughing your asterisks off? [15:25]
<gvainfo> how do I explain why pnestatus is funny [15:26]
<gvainfo> I think that they just take that I'm crazy looking at the screen and laugh [15:27]
<jeffg> gvainfo, are they far off from the truth of the matter? [15:27]
<gvainfo> probably not [15:28]
<jeffg> gvainfo, are they french? if so, explain that it's the same faux pas that mme de gaulle made when asked at a party in england what she was waiting for and replied, "happiness", but with her thick accent it came out "a penis" [15:30]
*p-brane makes no comment [15:30]
<Dalek3> anybody know what I'm missing? I've added the service, added a provisioning group with the same name as the service, and added it into the poller-configuration as a passive service [15:30]
<p-brane> Dalek3: did you import the provisioning group you created? [15:31]
<gvainfo> jeffg hmm [15:31]
<jeffg> Dalek3, i think you're mostly missing a comprehensive understanding of how the pieces of opennms work together. i know that doesn't help, but i have to point out that you seem to be in a hurry :) [15:31]
*** lisppaste7 :common-lisp.net joined the channel [15:31]
<Dalek3> yeah I hit the import button a few times.. it shows 0/0 nodes [15:32]
<gvainfo> jeffg no I have yet again to explain myself [15:32]
<gvainfo> you seem not to be a very good supporter [15:32]
<Dalek3> yeah I know I don't have a comprehensive understanding [15:32]
<Dalek3> I just installed it two days ago [15:33]
<gvainfo> Dalek3: did you add any node in the group? [15:34]
<Dalek3> no, that's what I'm trying to have happen [15:34]
<Dalek3> when a system sends a trap, I want the node, interface and service to be added [15:34]
<gvainfo> that's entirely different; the provisioning is used for _manual_ provisioning. [15:35]
<Dalek3> the trap is appearing in events and is formatted the way I configured in the event.xml but I don't understand how traps get associated with services [15:35]
<Dalek3> oh [15:35]
<Dalek3> I was asking about automatic creation of nodes when a trap was received [15:35]
<Dalek3> the provisioning groups looked to me like they were a manual process [15:36]
<Dalek3> and I don't want to have to manually add 1500 nodes [15:36]
<jeffg> Dalek3, the provisioning groups feature is a front-end [15:37]
<gvainfo> There might be more competent people here but I _think_ that the default config is that upon a received trap the node is "a new suspect" [15:37]
<Markw78> Manage Path Outages - gone from Admin screen in 1.5.9? [15:37]
<jeffg> look in OPENNMS_HOME/etc/*-imports.xml; you can programmatically generate that file and import it. [15:37]
<gvainfo> a "new suspect" is put into a capability scan. [15:37]
<Markw78> was trying to setup a path and couldn't find it anymore :/ [15:37]
<jeffg> Markw78, i think that moved to notifications [15:37]
<Markw78> ahh, ty ty [15:37]
<Dalek3> right, I was reading something about the LoopPlugin [15:37]
<Dalek3> will that do what I need? [15:38]
<p-brane> Dalek3: if it says 0/0 that means there are no nodes defined in the group you created [15:39]
<p-brane> Dalek3: sounds strange [15:39]
<Dalek3> http://www.opennms.org/index.php/Passive_Status_Keeper#Loop_Plugin_Method [15:40]
<Dalek3> I think the issue is that I'm starting out with this with the least documented part first [15:41]
<Dalek3> heh [15:41]
<Dalek3> I had added the service this last time through the psql command shown just below that [15:47]
<Dalek3> hope that doesn't break anything [15:51]
*** nmsmanjohn parted the channel [15:53]
<Dalek3> p-brane yeah there are no nodes definede [15:54]
<Dalek3> that's what I'm trying to have happen automatically [15:55]
<Dalek3> ahh! [15:57]
<Dalek3> found something in the trapd configuration [15:57]
<Dalek3> "new suspect on trap" is set to false [15:57]
<Dalek3> I'm going to enable that and see what happens [15:57]
<Dalek3> ahh! [16:01]
<Dalek3> now the trap is triggering a new suspect [16:02]
<Dalek3> it says its discovered a new interface and it's being queued for a services scan [16:02]
<Dalek3> now what is supposed to happen [16:04]
<Dalek3> heh [16:04]
<Dalek3> yay! the node has appeared! [16:05]
<Dalek3> it got auto created [16:05]
<Dalek3> now the only issue is that the node name is set to the interface IP [16:07]
<RangerRick> neat: http://hadoop.apache.org/core/ [16:08]
<RangerRick> looks like last.fm is using this for their backend [16:08]
<RangerRick> which is a lot less dirty than it sounds ;) [16:08]
<RangerRick> brozow: thanks for all the merge work! [16:12]
<brozow> no problem... almost done [16:13]
<RangerRick> yeah, I saw there were only a few left [16:13]
<brozow> I'm doing 2431 now.. and I think the rest of the defects are installation or something [16:13]
<RangerRick> and 2428 [16:14]
<RangerRick> but other than that, yeah [16:14]
<brozow> but there are a few revisions still pending too that we should investigate [16:14]
<RangerRick> could probably just verify 1098 also [16:14]
<RangerRick> yeah, I was gonna wait for everything to get merged before I did another run of "what's not got a bug" [16:14]
<brozow> oh yeah I missed 2428.. I'll do that one now [16:15]
<RangerRick> haha, "Companies like Hyperic and Zenoss want to be 100+ million companies in 5 years or less. Unless I am completely off the wall, I can’t see anyway OpenNMS w/o VC funding can do something like that." [16:16]
<RangerRick> and I can't see how Hyperic and Zenoss are gonna do it either, but it's not worth responding at this point ;) [16:16]
<brozow> heh.. that John Willis? [16:18]
<RangerRick> yeah [16:18]
<RangerRick> http://www.johnmwillis.com/nagios/why-only-two/ [16:18]
<RangerRick> his whole argument is working on the assumption that they'll succeed, and mine is on the assumption that they will fail, that's the ultimate issue [16:20]
<brozow> yeah... I can see that... [16:20]
<RangerRick> I've seen too many good VC-funded companies go down to expect some mediocre ones to succeed, but hey, if they get lucky, more power to them. In the end, we're going after completely different markets. [16:20]
<brozow> I'm pretty certain Hyperic will fail [16:21]
<RangerRick> OpenNMS targets grassroots in-the-trenches "hey, we use this stuff because <other_vendor> isn't doing what we need, and we can tweak this open-source stuff" [16:21]
<brozow> I'm not as confident about Zenoss [16:21]
<RangerRick> not "we're going to spam marketing until enough CTOs know about us and force their IT teams to use us" [16:21]
<brozow> heh [16:22]
<RangerRick> it's a retarded way to do business, despite the fact that it's the most common model nowadays [16:22]
<brozow> yeah I agree [16:22]
<brozow> but its a a good way to get bought [16:22]
<Markw78> Does NMS use PTR DNS records to find node names rather then interface IP's? [16:22]
<RangerRick> which just spreads the money around [16:22]
<brozow> I don't see 1098 on the 1.5.91 Milestone.. is it somewhere else? [16:23]
<RangerRick> hm, it is: http://bugzilla.opennms.org/show_bug.cgi?id=1098 [16:23]
<RangerRick> you still using the bug deps? or target milestones? [16:23]
<brozow> the target milestone... looking in bug deps now [16:24]
<RangerRick> 'cause it showed up in milestones for me [16:24]
<brozow> hmmm... how do you get your query [16:24]
<brozow> http://bugzilla.opennms.org/buglist.cgi?action=wrap&bug_file_loc=&bug_file_loc_type=allwordssubstr&bug_id=&bugidtype=include&chfieldfrom=&chfieldto=Now&chfieldvalue=&email1=&email2=&emailassigned_to1=1&emailassigned_to2=1&emailcc2=1&emailreporter2=1&emailtype1=substring&emailtype2=substring&field0-0-0=noop&long_desc=&long_desc_type=substring&short_desc=&short_desc_type=allwordssubstr&status_whiteboard=&status_whiteboard_type=allwordssu [16:24]
<brozow> bstr&type0-0-0=noop&value0-0-0=&votes=&=&target_milestone=1.5.91%20%281.6%20beta%202%29&bug_status=RESOLVED [16:24]
<Markw78> thats a hell of a url [16:25]
<brozow> hmm... that was bad... sorry [16:25]
<brozow> RangerRick: is it ResolveD? [16:26]
<brozow> ah no.. its closed [16:26]
<RangerRick> http://bugzilla.opennms.org/buglist.cgi?query_format=advanced&target_milestone=1.5.91+%281.6+beta+2%29&bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&bug_status=RESOLVED&bug_status=CLOSED&resolution=FIXED&cmdtype=doit [16:26]
<RangerRick> ah, yeah, that would be it [16:27]
<RangerRick> that's the query I use [16:27]
<RangerRick> (after trimming out the pintless stuff) [16:27]
<RangerRick> *pointless [16:27]
<Markw78> will my poller log show the HTTP status it received from an HTTP poll? [16:27]
<Markw78> guess I can go look ;o [16:27]
<brozow> I was just searching for 'Resolved' for 1.5.91 [16:27]
<RangerRick> that would be it [16:28]
<brozow> i didn't think there was a revision for 1098 so I just closed [16:28]
<RangerRick> right [16:29]
<RangerRick> technically it probably should have just been marked as a duplicate of something else related t othe jrobin stuff I did at dev-jam, but... meh [16:29]
<brozow> ok [16:29]
<Dalek3> now the only thing I need is a way to automatically use the SNMP community name as part of the node name [16:30]
<brozow> do any of the remaining resolveds require merges? [16:30]
<RangerRick> so what's left is... r8966,8971-8972,8988 -- a bunch of event stuff that needs to be OKd [16:30]
<RangerRick> and r9005 which is dj's PSM fix [16:31]
<brozow> ok.. I'll look at 9005 [16:31]
<RangerRick> and then 9031 which is my release notes update [16:31]
<RangerRick> which I'll merge right now [16:31]
<brozow> ok [16:31]
<RangerRick> and then I have ot run, we're going to the This American Life thing at North Hills tonight =) [16:31]
<RangerRick> nationwide "live" broadcast thing, kinda interesting [16:32]
<RangerRick> ok, mine's committed, if you want to update and merge/check 9005 [16:32]
<brozow> ok.. is there a bug for 9005? [16:34]
<Markw78> someone needs to write a filter app for nms logs :P [16:34]
<RangerRick> brozow: I'm not sure [16:35]
<RangerRick> I thought there might be [16:35]
<RangerRick> 2438 [16:36]
<RangerRick> oh, it was dave's [16:36]
<RangerRick> not dj [16:36]
<brozow> ok 9005 is commited [16:39]
<RangerRick> cool [16:39]
*RangerRick heads out [16:40]
<brozow> thanks for the hard work RangerRick! have a fun time! [16:40]
<RangerRick> will do, thank you as well, another solid release on the way :) [16:40]
<RangerRick> the bug thing is actually working out reasonably well [16:41]
<brozow> I agree [16:41]
<RangerRick> and it's going to be more useful when we start truly managing two releases [16:41]
<brozow> yeah that's for sure [16:41]
<brozow> Markw78: when do you think you can have a prototype? [16:41]
<RangerRick> but we've gotta get this one out first ;) [16:41]
<brozow> yeah.. boy won't that be great! [16:42]
<RangerRick> I wonder if I should go ahead and start the tag [16:44]
<RangerRick> if we have to do any fixes, we can hand-merge them [16:45]
<RangerRick> but this should be it [16:45]
<brozow> good question... did that event stuff get merged? [16:45]
<RangerRick> hm, no it didn't [16:45]
<RangerRick> perhaps wait, we can check on those tomorrow [16:45]
<brozow> who should do that do you think? [16:45]
<brozow> I suppose I can look at them [16:46]
<RangerRick> any of us can, I guess, just need to make sure things start :) [16:46]
<brozow> alright I'll do it then [16:46]
<brozow> do they have bugs ? [16:47]
<RangerRick> not that I could find [16:49]
<brozow> yeah.. I can't find them either [16:49]
<RangerRick> svnmerge.py avail -S 1.6-testing -l [16:50]
<RangerRick> will give you the logs, btw [16:50]
<brozow> ah ok [16:50]
<Markw78> 10000ms timeout on a poller is 10 seconds right ;x [16:50]
<brozow> yeah its also a LONG time [16:52]
*sartin follows RangerRick's link. [16:53]
<sartin> who is johnmwillis? [16:53]
<sartin> ... sounds like a Tobacco company. [16:53]
<brozow> heh [16:55]
*brozow waits for remaining revisions to build [16:55]
<sartin> brozow: I was asking about ticketer plugins earlier on... [16:56]
<brozow> oh yeah? anything fun? [16:57]
<Markw78> I need to get the Jira plugin working [16:57]
<Markw78> hope thats not as much of a pain as I am anticipating [16:57]
<sartin> I have one nearly down for OTRS, should it be an add-on or part of the core distribution? [16:58]
<brozow> you can make it part of the normal code if it doesn't depend on anything proprietary [16:59]
<brozow> jira and centric both needed libs not available in maven so that's why they are under tools [16:59]
<sartin> No, it's all GPL. [16:59]
<brozow> so no problem then [17:00]
<sartin> OK, I'd assumed it was licencing, but wanted to check. [17:00]
<brozow> what language is it written in? [17:01]
<sartin> The OpenNMS end is java/axis, the OTRS end is Perl/SOAP::Lite [17:01]
<brozow> ah ok [17:01]
<sartin> I'm hoping the OTRS guys will host my OTRS module, but I've only just started a discussion with them. If not we'll need to find a home for it. [17:02]
<brozow> ok [17:02]
<brozow> RangerRick is probably the best guy to ask about that [17:03]
<sartin> OK, It's not urgent. I can start work on a feature branch next week. Code is done and working, but I'm out of time for now. [17:04]
<sartin> Nothing that needs to get in the way of the next release. [17:05]
<brozow> ok [17:05]
<brozow> thanks for the hard work! [17:05]
*sartin grins [17:05]
<sartin> I blogged about it today. [17:05]
<sartin> brozow: thinking about it, a feature branch after the next release would make sense. [17:08]
<brozow> cool [17:08]
<sartin> brozow: OK thanks for that, off to get some sleep. Catch you next week. [17:12]
*** MrCerulean parted the channel [17:12]
<brozow> good chattin' with you sartin... tell your oldest happy b'day for me [17:12]
<sartin> will do, thanks mate... [17:13]
*** mhuot :pdpc/supporter/active/mhuot joined the channel [18:15]
*** moisdtech :216.120.161.118 joined the channel [18:45]
*** moisdtech parted the channel [18:46]
*** jeffg :pdpc/supporter/active/jeffg joined the channel [19:11]
*** PowerOp :116.59.44.10 joined the channel [20:29]
*** jeffg :pdpc/supporter/active/jeffg joined the channel [20:41]
<jeffg> oof, another incredibly unfortunate name: http://www.engadget.com/2007/08/24/trekstor-showcases-i-beat-blaxx-portable-media-player [20:48]
<djgregor> heh [20:49]
<djgregor> out of frame? [20:49]
<djgregor> ;-) [20:49]
<jeffg> djgregor! [20:50]
<djgregor> jeffg! [20:50]
<jeffg> djgregor, ever worked with cisco mars? [20:50]
<djgregor> no [20:50]
<djgregor> what is it? [20:50]
<mhuot> Logging server [20:50]
<djgregor> is it part of their IDS system? [20:51]
*djgregor watches the video for Prodigy's Smack My Bitch Up for the first time in many years [20:51]
<mhuot> Somewhat [20:51]
<jeffg> sorta [20:51]
<jeffg> it's an aggregator of security-related events [20:51]
<djgregor> ah [20:51]
<djgregor> I think we use that at $WORK [20:51]
<jeffg> eats all manner of snmp traps and syslogs [20:52]
<jeffg> so one would think it would *generate* decent traps [20:52]
<chalco> jeffg: you are clearly asking too much [20:52]
<jeffg> but the trap (yes, singular) it sends is complete bunk [20:52]
<mhuot> heh [20:52]
<mhuot> lol "so one would think it would *generate* decent traps" [20:52]
<jeffg> two required varbinds, both displaystrings [20:53]
<jeffg> "MARS-1-101" is the first (hard-wired) varbind. the second is just the *raw* syslog datagram regurgitated into a displaystring. like, with the angle-bracketed facility + severity and everything. [20:54]
<djgregor> wow [20:54]
<jeffg> i actually punted on doing anything with this thing's traps and opted to use syslog instead [20:54]
<djgregor> it's a lot diferent than I remember [20:54]
<jeffg> 'cause you can at least mine the syslog messages for useful info if you use syslogd [20:54]
<jeffg> needless to say, cisco bought it [20:55]
<fastjay> mars gah! [20:56]
<fastjay> <-- not a fan [20:56]
<jeffg> mars attacks [20:56]
<_sndbot> calorie intake management with balanced nutrition [20:56]
<fastjay> though it does make some pretty graphs [20:56]
<djgregor> 4.4k line source files are teh suck [20:56]
<_sndbot> perhaps wait, we can tweak this open-source stuff" [20:56]
<jeffg> fastjay! [20:56]
<fastjay> jeffg! [20:57]
<fastjay> how goes it? [20:57]
<chalco> I could market crap on a stick and people would buy it if it made pretty graphs [20:57]
<jeffg> goes well. you? [20:57]
<fastjay> not to bad [20:57]
<fastjay> doing some dirty things to make our provisioning system clear puppet certificates when a box reinstalls [20:57]
<jeffg> chalco, i suppose you could draw graphs using a poo-stick [20:58]
<jeffg> heh [20:58]
<chalco> there's an idea [20:58]
<fastjay> going ot have to port it into puppet when they move from XML-RPC to rest [20:58]
<fastjay> cause xml-rpc w/ wget sounds kinda painful [20:58]
<mhuot> I swear I saw some rrd graphs in some product demo the other day, but I could not see the rrdtool tag on them [20:58]
<djgregor> jeffg: I didn't catch the name of that device until I just read it out lout to Sharon [20:59]
<mhuot> Ah, now I remember it was on Juniper's SSL vpn tool [20:59]
<jeffg> djgregor, lulz [20:59]
<jeffg> ~seen joed [20:59]
<_sndbot> jeffg, At 4/29/08 12:31 PM joed said: Or do you have error messages around that collection? [20:59]
<jeffg> damn, what is he, moving house or something? [21:00]
<fastjay> he is supposed to be in my area in a few days [21:00]
<fastjay> but not sure when [21:00]
<fastjay> i wanna meet up w/ home for dinner or drinks or something [21:00]
<jeffg> fastjay, beware drinking with joed [21:00]
<jeffg> but you knew that :) [21:00]
<fastjay> oh i know all about that [21:00]
<fastjay> we found a bar one night at dev jam [21:00]
<fastjay> all i know is.. we couldn not find the hall [21:00]
<fastjay> we wondered all around campus [21:01]
<jeffg> did the third guy duck? [21:01]
<fastjay> ;) [21:01]
<jeffg> at the client i'm with this week, they put up financial newspapers over the urinals. one of them that was up today has a picture of a russian banker who is a dead ringer for a slightly annoyed joed. very disconcerting to take a leak with that staring you in the face. [21:02]
<jeffg> i googled the guy's name and fortunately he looks nothing like joed, it's just that one pic [21:02]
<mhuot> slightly annoyed joed...isn't that his default state? [21:03]
<jeffg> mhuot, yeah, i'm into stating the obvious today [21:03]
<mhuot> heh [21:03]
<jeffg> so one would think it would *generate* decent traps [21:03]
<_sndbot> no biggy for him [21:03]
<mhuot> I can't name 5 devices that give decent traps [21:03]
*jeffg resists the temptation to try that exercise [21:04]
<mhuot> I dare ya! [21:04]
<mhuot> Times up! [21:05]
<mhuot> :-D [21:05]
<jeffg> hmph [21:05]
<jeffg> dell sure as hell ain't on the list [21:06]
<mhuot> lol [21:06]
<mhuot> They all suck [21:06]
<djgregor> ur mom [21:06]
<fastjay> whats worng w/ dell traps? [21:06]
<mhuot> Just varying degrees [21:06]
<mhuot> djgregor: :-P [21:07]
<jeffg> fastjay, too many different mibs, each of which is just different enough in its conventions to be really annoying when you have a mix [21:07]
<djgregor> mhuot: evening. :-) [21:07]
<mhuot> ~pet djgregor [21:08]
*_sndbot snuggles with djgregor and ruffles djgregor's hair and then pulls the thong. [21:08]
<mhuot> Oh!! And it is THONG THURSDAY EVEN!! [21:08]
<djgregor> OH YEAH! [21:09]
<jeffg> ~dance [21:10]
<_sndbot> jeffg, I have no idea what dance is. [21:10]
<chalco> ~w00t [21:11]
<_sndbot> \o/\o/\o/\o/ [21:11]
<fastjay> jeffg: well there is that [21:13]
<fastjay> jeffg: i've found that omsa (hardware) and then omsa storage are the two bits you need [21:13]
<fastjay> when i was at fast i had a generic omsa poller you fed a chunk of html.. and it would monitor all sorts of omsa related stuff [21:14]
<fastjay> s/html/xml/ [21:14]
<jeffg> ~dance [21:14]
*_sndbot dances three-quarters naked in time with the music playing in his little bot-brain [21:14]
*** HTI :142.179.92.143 joined the channel [21:14]
*** HTI parted the channel [21:15]
<chalco> so how did things jump from 1.3.11 to 1.5.90? [21:17]
<chalco> getting a little loose with version numbers are we? [21:17]
<jeffg> chalco, we thought that 1.3 -> 1.4 didn't fully represent the progress we've made [21:17]
*chalco should try it out [21:17]
<chalco> sure, but .90? [21:17]
<jeffg> think of it as 1.6-beta1 [21:18]
<chalco> I see [21:18]
<jeffg> so stable will jump 1.2 -> 1.6 [21:19]
<chalco> that fits the way java versioning jumped around :D [21:19]
<jeffg> sun's really good at that stuff [21:19]
<jeffg> they can even start numbering at 2.0 and retroactively rename sunOS 4.x to solaris 1.x [21:20]
<djgregor> ha ha ha [21:21]
<chalco> ~dance [21:22]
*_sndbot dances totally bare-assed naked in the street [21:22]
*chalco is feeling fine tonight [21:22]
<djgregor> since we have a few people around, I have a question about the user security part of the webUI that is one of the summer's GSOC projects (limiting users to a view of everything in OpenNMS) [21:24]
<djgregor> how should we define a view? a filter (like we do in the poller, collectd, etc.)? [21:24]
<jeffg> djgregor, i've imagined it as something aopish (to the extent my tiny brain can get itself around aop) [21:25]
<djgregor> yeah, but that just tells us how we code it, right? I'm wondering how we actually configure the views for the code to use [21:25]
<mhuot> I'd like us to use a filter, but hopefully reusing a filer [21:27]
<Cuchulain> hrm [21:27]
<mhuot> (s/filer/filter/) correction: I'd like us to use a filter, but hopefully reusing a filter [21:27]
<jeffg> filter's good, but applies only to interfaces as currently defined [21:27]
<djgregor> jeffg: yeah, I'm thinking a filter would only work well if we can make that more flexible [21:27]
<jeffg> whatever defines a view needs to be applicable to any OnmsResource, right? [21:28]
<mhuot> onmsnode ? [21:28]
<djgregor> it needs to be more flexible for it to truely sanely work for some of the current uses, even [21:28]
<djgregor> hmm [21:28]
*djgregor looks at the children of OnmsResource [21:28]
<Cuchulain> what's the intention of the user security part? [21:28]
<djgregor> Cuchulain: limit user views into the OpenNMS interface [21:29]
<djgregor> e.g.: so people can let customers login and only see their own stuff [21:29]
<Cuchulain> down to server/interface level? [21:29]
<Cuchulain> awesome [21:29]
<Cuchulain> I need this :) [21:29]
<Cuchulain> er, service/interface level [21:29]
<djgregor> Yeah, if I'm an ISP, it would be great to let my customer see *both* ends of the link, I think [21:29]
<Cuchulain> yeah [21:30]
<djgregor> jeffg: OnmsEntity, probably, not OnmsResource (although it's related to an entity) [21:30]
*jeffg looks [21:31]
<djgregor> Heh, OnmsEvent is a child of that as is OnmsVulnerability [21:31]
<djgregor> maybe not all entities, then. ;-) [21:31]
<jeffg> well, OnmsEvent can in theory be viewed in the web ui, once we make the event viewer a proper view [21:35]
<djgregor> sorry, I'm thinking about this from a different angle: would we ever write a filter that says a user can only see certain events (vs. letting them see events on certain nodes, interfaces, services, etc.)? [21:36]
<mhuot> hmmm [21:36]
<djgregor> so, if we set a filter that says only let a user see node 2 and interface 94 (on some other node), then that would filter everything in the UI, including events, graphs, etc.. [21:37]
<djgregor> but would we set a filter that said a user can only see event 491? [21:37]
<djgregor> or vulnerability 941 [21:37]
<mhuot> The more I think about it no [21:37]
<djgregor> i.e.: I'm thinking we need to clearly define filterable entities, which are probably nodes, IP interfaces, SNMP interfaces, services, maybe service types (HTTP) [21:38]
<djgregor> and there are lots of ways to select them that filters give you today [21:39]
<jeffg> sounds like a job for an interface [21:39]
<djgregor> If filters can be made to not be specific to an IP interface, we have FilterDao. :-) [21:40]
*djgregor isn't convinced filters are the best way to go, but can't think of a better way, and they sure are a very opennmsie way to do things, which is good [21:40]
<jeffg> hmm... what about events that pertain to a node that is shared by several customers, but that should be seen by no customers? e.g. cisco config mgmt trap events? [21:41]
<djgregor> hmmm... good one [21:41]
*djgregor is glad he is logging. :-) [21:41]
<djgregor> jeffg: if the filter doesn't match the node (just an interface on the node), then you wouldn't see anything except events specific to that interface [21:42]
<djgregor> but filtering out certain sets of data seems like a good idea [21:42]
<djgregor> like maybe no graphs [21:42]
<djgregor> no events [21:42]
<djgregor> etc. [21:42]
<jeffg> given a router on which both coke and pepsi have interfaces, they should both be able to see e.g. memory pool graphs and maybe even CPUHOG syslog events, but not the results of your router jockey's config changes on behalf of dr pepper [21:44]
<Cuchulain> sounds good. [21:45]
<Cuchulain> all done by next week, right? :) [21:46]
<jeffg> geez, i gotta save all your asses. if i help, we can do it in five minutes! [21:46]
<_sndbot> if I read the whole wiki-page..... You'll like scandinavia [21:46]
<djgregor> ha ha [21:46]
<jeffg> <3 hackers [21:47]
<jeffg> btw, has anybody else picked up on the fact that jonny lee miller is eli stone? [21:54]
*** exothermc_ :68-189-133-163.dhcp.wlwl.wa.charter.com joined the channel [22:24]
<RangerRick> am I suppose to know who either of those names are? =) [22:28]
<fastjay> wow.. when did it become 8:40? [22:32]
<fastjay> and why am i still @ the office ? [22:32]
*** PowerOp_ :134.137-60-66.DIA-subnet.surewest.net joined the channel [22:49]
*** ed_mann :c-98-214-139-110.hsd1.il.comcast.net joined the channel [22:50]
<_snd> morn babes [23:18]
*** exothermc_ :68-189-133-163.dhcp.wlwl.wa.charter.com joined the channel [23:28]