Lefthand OIDs still not working
|
|
|
||||||||
|
|
||||||||
|
|
On Wed, Nov 4, 2009 at 23:44, Jeff Gehlbach <[hidden email]> wrote:
> On Nov 4, 2009, at 12:40 PM, Dennis Crandall wrote: > >> 1) Created a new snmp-collection section in the >> datacollection-config.xml file. Its name is: Lefthand >> 2) Added the appropriate graph entries into snmp-graph >> 3) Created a new collectd-configuration section in >> collectd-configuration.xml file > > Dennis, this is all good information, thanks for providing it. Could > you reply with your collectd-configuration.xml and datacollection- > config.xml in their entirety, perhaps zipped up to keep the size > down? I know you posted snippets of these before, but seeing them > whole and with any changes you've made since the last go-round can > make all the difference. Also helpful would be a screenshot of the > node detail page for the Lefthand node in the OpenNMS web UI. > > I notice that this post isn't threaded under your previous messages on > this topic. There's a classic article with some excellent tips on, > among other things, maintaining threading on mailing lists as a way to > get your questions answered on lists like this one: > > http://catb.org/~esr/faqs/smart-questions.html > > -jeff I have attached the files. Thanks! Dennis ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Please read the OpenNMS Mailing List FAQ: http://www.opennms.org/index.php/Mailing_List_FAQ opennms-discuss mailing list To *unsubscribe* or change your subscription options, see the bottom of this page: https://lists.sourceforge.net/lists/listinfo/opennms-discuss  LefthandConfFile.tgz (125K) Download Attachment LefthandSanChooseGraphicsSnapshot.jpg.gz (27K) Download Attachment LefthandSanUISnapshot.jpg.gz (98K) Download Attachment |
||||||||
|
|
|
||||||||
|
|
|
||||||||
|
|
On Fri, Nov 6, 2009 at 10:10, Dennis Crandall <[hidden email]> wrote:
> On Thu, Nov 5, 2009 at 23:23, Jeff Gehlbach <[hidden email]> wrote: >> On Nov 5, 2009, at 8:24 AM, Dennis Crandall wrote: >> >>> I have attached the files. >> >> OK, from the node detail page screenshot I see that the ATXDEVSAN01's >> sysObjectID is .1.3.6.1.4.1.8072.3.2.10. As I think you know already, >> that means "a Net-SNMP agent running on Linux", so you've created a >> custom SNMP collection in datacollection-config.xml called "Lefthand" >> and made a new package in collectd-configuration.xml that includes >> only the ATXDEVSAN01 node's primary interface. One technicality >> that's probably not an issue: you used a "sysoidMask" in the systemDef >> you made in the new collection, but when you know the exact >> sysObjectID you should use "sysoid" instead since "sysoidMask" checks >> for a prefix match rather than an exact match. Also, you mentioned >> you'd tried changing the "name" attribute of your systemDef; for the >> record, there's nothing magical about the name. You could have called >> it "Jim" or "Glenn" if you wanted. >> >> You created a set of custom resourceTypes in your new collection for >> the objects in the Lefthand MIB, but you never actually use these >> types -- normally the name of the appropriate resourceType would >> appear as the value of the "instance" attribute for all the mibObj >> elements in one group. Instead, you have specified constant numeric >> instance IDs in all your mibObjs. That's OK, but it's not doing what >> you wanted. >> >> I think what you need to do is this. In your "lhn-modules" collection >> group, change the "clusModStorStatus1" mibObj so that its alias is >> just "clusModStorStatus" and its instance is "clusModuleEntry" (the >> name of the resourceType you defined to represent a module in a >> cluster). Kill the "clusModStorStatus2", "clusModStorStatus3", and >> "clusModStorStatus4" mibObjs. Likewise, kill mibObjs >> "clusModStatsIOsRd2" through "clusModStatsIOsRd4" and change mibObj >> "clusModStatsIOsRd1" so that its alias is just "clusModStatsIOsRd" and >> its instance is "clusModuleEntry". Do the same for each of the other >> "foo1-4" mibObjs in this group. Finally, track down the OID for the >> "clusModuleName" object and add a mibObj for it to this group, with >> the same instance. Without doing this last item, your resource labels >> will read literally "${clusModuleName} (index N)". >> >> Also, since you're using OpenNMS 1.6.5, you need to add one more group >> to your "Lefthand" collection in order to work around bug 2447: >> >> <group name="bug-2447-workaround" ifType="all"> >> <mibObj oid=".1.3.6.1.2.1.2.2.1.2" instance="ifIndex" >> alias="obIfDescr" type="string" /> >> </group> >> >> You need not reference this group anywhere; it just needs to be >> present inside your "Lefthand" collection. After you upgrade to 1.6.6 >> or later you can remove it. >> >> At this point I would suggest restarting OpenNMS so that the changes >> to datacollection-config.xml will take effect. Wait 10 minutes and >> then enter the resource graphs workflow for the ATXDEVSAN01 node. If >> you've gotten everything right, you should now see a new pick-list in >> the "Choose Resource Graphs" page with a header of "LHN Cluster Module >> Name". There should be one entry in the list for each cluster module >> in the SAN. The graphs will still be broken, but the presence of this >> list is your indication that you've straightened out one of your >> collection groups. >> >> If that worked, you'll move on to remediate the "lhn-cluster" and "lhn- >> volume" groups in the same style as you did the "lhn-modules" group >> above. I believe you'll have nine mibObjs in "lhn-cluster" (all with >> instance="clusClusterEntry") after you add the string one for >> clusClusterName. And there should be ten mibObjs in lhn-volume, all >> with instance="clusVolumeEntry" once you've added the string one for >> clusVolumeName. >> >> The last thing will be to remediate the graph definitions. As a >> starter, your lefthand.modulestatsIOsRead definition should get the >> following changes: >> >> report.lefthand.modulestatsIOsRead.columns=clusModStatsIOsRd >> report.lefthand.modulestatsIOsRead.type=clusModuleEntry >> >> And instead of four DEFs, four LINE2s, and twelve GPRINTs, you'll have >> one DEF: >> >> DEF:clusModStatsIOsRd={rrd1}:clusModuleStatsIOsRd:AVERAGE >> >> And you'll use that clusModStatsIOsRd token in the one LINE2 and three >> GPRINT statements. >> >> I hope this helps. I know it's a lot to digest. >> >> -jeff no graphs. I have attached the apropos sections of datacollect and snmp-graph files plus screenshots. Thanks! Dennis ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Please read the OpenNMS Mailing List FAQ: http://www.opennms.org/index.php/Mailing_List_FAQ opennms-discuss mailing list To *unsubscribe* or change your subscription options, see the bottom of this page: https://lists.sourceforge.net/lists/listinfo/opennms-discuss |
||||||||
|
|
|
||||||||
| Free Embeddable Forum Powered by Nabble | Help |
