Am 25.03.2013 16:50, schrieb Mario Gzuk:
Am 25.03.2013 16:24, schrieb Gonéri Le Bouder:
On Mon, Mar 25, 2013 at 03:26:49PM +0100, Mario Gzuk wrote:
Hopefully the last question about that task. If I run the
fusioninventory-esx
directly I got an correct deviceid for each host:
<DEVICEID>esx1-2013-02-11-14-23-51</DEVICEID>
<DEVICEID>esx2-2013-02-11-14-25-32</DEVICEID>
If I run this through the agent/rest-esx the inventory xml data
contains the
deviceid from the host I am running the task (for every esx host the
same):
<DEVICEID>a-test-2013-03-21-16-54-58</DEVICEID>
Hi Mario,
Actually, none of them are correct. The deviceid is a string stored
on the
computer. It is used to identify the agent. This string is:
- the hostname
- the installation date.
We can store the string on a ESX server. In order to reduce
compatibility
issue we decide to generate a “fake” string every time.
This deviceid will also be lost for a common computer is the agent is
removed (and the “var” dir lost).
Best regards,
--
Gonéri Le Bouder
So why there is no "fake" deviceid generated for each esx host when I
run this through the agent? It differs when I run:
perl fusioninventory-esx --host bla ->
<DEVICEID>bla-2013-02-11-14-23-51</DEVICEID>
perl fusioninventory-agent (it gets the jobs via rest-esx) the xml
gets the <DEVICEID>a-test-2013-03-21-16-54-58</DEVICEID>
both commands are running on the same host.
This is a little bit confusing to me, because I try to manage the data
to the device id -> if a new one is generated a new set of data is
generated. But in this case I have got 3 different inventory data (2
from esx hosts and one from the host itself) to the same device id...
On which identifier you check that? The host name? Or is the
hardware:UUID every time the same?
Thank you for your time to answer my questions...
Cheers Mario
So now I think I can describe my problem a little bit mor in detail. I
have different inventory sets, e.g. 2 from ESX and 2 from the host it
self (one from a linux installation and one from a windows
installation), all has the same deviceids (maybe the linux installation
has its own). At this state all is fine, bit if the deploy task is
asking for his config/jobs the deploy task is asking only with the
deviceid. Now I have 3 deviceids with different hosts behind, how to
decide which jobs has to be run for which host? This isnt possible @this
state with the information the deploy task is asking.
Do you know what I mean? Maybe it is hard when a german try to explain
some what in english for also not native english speaking people... So
feel free to ask again if you think I am writing only opaque things...
Cheers
_______________________________________________
Fusioninventory-devel mailing list
[email protected]
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/fusioninventory-devel
