[dev.icinga.com #674] IPv6 and Multiple adresses for one host #338
Comments
Updated by mfriedrich on 2010-09-02 10:38:09 +00:00
this matches IPv6 too. http://feedback.icinga.org/forums/50329-general/suggestions/890661-ipv6?ref=title
|
Updated by mfriedrich on 2011-01-17 09:05:27 +00:00
|
Updated by mfriedrich on 2011-01-17 14:49:51 +00:00 http://www.thibault.info/aggregator/categories/2?page=2
|
Updated by mfriedrich on 2011-01-24 21:33:01 +00:00
would be interesting for the next unstable branch. if changing the objects.h it is necessary to put the newly added object at the end of the struct (typecasting and adress mapping, otherwise things may segfault the way they are being used (mklivestatus!)). |
Updated by mfriedrich on 2011-02-07 17:28:17 +00:00
|
Updated by mfriedrich on 2011-02-07 17:33:13 +00:00
with applied patch using the above config, it looks like this from the logs (having a wrong path for check_ping too)
and on the extinfo.cgi it looks like this - ipv6 is only shown if defined (that's a minor change i did on the original patch, the comma would look strange in case) |
Updated by mfriedrich on 2011-02-07 17:47:54 +00:00 if used with icinga-web, it needs aptched idomod/ido2db too and updated database schema. idomod.c: this requires a change in the way data is being sent on the socket, and then being put into the database (config queries). but as a matter of fact, i think this is mandatory to complete this feature ;D |
Updated by mfriedrich on 2011-02-08 11:38:11 +00:00 had some issues bringing ido2db into shape (IDO_MAX_DATATYPES is an array size identifier, if not incrementing when adding a new data_type, accessing a wrong index in the array (size-1+1) will segfault then. but right on, address6 column on icinga_hosts is being populated too. so icinga-web could select that too in further implementations. will push to my branch after some more testing. |
Updated by mfriedrich on 2011-02-08 14:59:16 +00:00
from objects.cache - everything works as expected.
|
Updated by mfriedrich on 2011-02-08 18:17:47 +00:00 if address6 is not set, the host name is taken instead (for dns resolution purposes). the classic ui negotiates this by comparing if name==address6 and in that case not showing address6 as 'real' ipv6 address. neither should do icinga-web then - the database holds the default hostname for the address6 if nothing set. this is on purpose for other readings too (like the core behaves in objects.cache). |
Updated by mfriedrich on 2011-02-09 17:06:27 +00:00
check icinga-web issue #1206 ready for release. maybe add some docs hints somewhere. |
Updated by calestyo on 2012-08-28 15:44:52 +00:00 Stupid question: Wasn't all that already possible via custom object variables? |
Updated by mfriedrich on 2012-08-28 16:15:43 +00:00 sure, but which gui did have native support for custom variables? likewise, a native macro, other than a custom var macro, is even better. it might be a small change, but for productive usage, i prefer something not "volatile" in many cases. |
This issue has been migrated from Redmine: https://dev.icinga.com/issues/674
Created by mfriedrich on 2010-08-04 07:10:28 +00:00
Assignee: mfriedrich
Status: Resolved (closed on 2011-02-09 17:06:27 +00:00)
Target Version: 1.3
Last Update: 2012-08-28 16:15:43 +00:00 (in Redmine)
http://feedback.icinga.org/forums/50329-general/suggestions/890667-multiple-adresses-for-one-host?ref=comments
Some hosts today have more then one interface, and the common way to monitor these would be to create a host object for EACH interface based on the fact that each interface has it's own address.
What about making it possible to give several adresses to each host, so it all appears in the same ... more
Some hosts today have more then one interface, and the common way to monitor these would be to create a host object for EACH interface based on the fact that each interface has it's own address.
What about making it possible to give several adresses to each host, so it all appears in the same host object?
It would require some logic regarding what check to run on what interface, but it can be solved!
Attachments
Changesets
2011-02-08 17:01:06 +00:00 by mfriedrich de9bc5a
Relations:
The text was updated successfully, but these errors were encountered: