Il 07/04/2011 09:13, Pablo Fernandez ha scritto:
> Thanks Alessandro,
>
> I understand this procedure is done in a local nagios instance, right? (hence
> tests are not using ops) Or did you organize it with your Tier1?
Hi Pablo,
you can active the monitoring of uncertified sites also on the official
nagios; however, if you prefer using a VO different from ops you might
need another nagios server
cheers,
Alessandro
> BR/Pablo
>
> On Wednesday 06 April 2011 14:53:24 Alessandro Paolini wrote:
>> Il 06/04/2011 14:04, Pablo Fernandez ha scritto:
>>> Hi,
>>>
>>> Thanks for your answers.
>>>
>>> The idea was to create a permanent pre-production cluster, not really to
>>> check a single separate node... but it's good to know just in case one
>>> needs to add a new node or make a temporal test.
>>>
>>> So maybe the best is what Alessandro Paolini suggested, create a "tuned"
>>> nagios instance, a top-bdii and a WMS to be used by it... but then you
>>> don't have an ops certificate to launch the tests.
>>>
>>> How about creating a test/uncertified site? Could this be reasonable?
>>>
>>> BR/Pablo
>> hi Pablo,
>> a colleague of mine made me notice that the procedure for monitoring the
>> uncertified sites has been modified with the update 09:
>> https://tomtools.cern.ch/confluence/display/SAMDOC/Update-09
>>
>> now it is managed by yaim
>>
>> Cheers,
>> Alessandro
>>
>>> On Wednesday 06 April 2011 13:40:05 Gonçalo Borges wrote:
>>>> Hi Tiziana...
>>>>
>>>> Unfortunately, this setup will still generate alarms in the dashboard,
>>>> and ROD will have to act accordingly. We already reported this as an
>>>> OTAG requirement.
>>>>
>>>> SO, the best option is to put a downtime in the node, otherwise tickets
>>>> will start to be open.
>>>>
>>>> Cheers
>>>> Goncalo
>>>>
>>>> On 04/05/2011 06:20 PM, Tiziana Ferrari wrote:
>>>>> Hi.
>>>>>
>>>>> If it's just a node that you need to test, you can flag it ‘monitored’
>>>>> and non production in GOCDB.
>>>>>
>>>>> Tiziana
>>>>>
>>>>> On 05/04/2011 16:59, Pablo Fernandez wrote:
>>>>>> Dear all,
>>>>>>
>>>>>> Does anybody know if there's an official way to create and monitor
>>>>>> (without
>>>>>> accounting, tickets or avail reports) a pre-production cluster?
>>>>>>
>>>>>> Of course you can always create your own top-bdii and include your
>>>>>> cluster
>>>>>> there, but this will not include gLite Nagios tests, which is always
>>>>>> nice to
>>>>>> have in such an environment.
>>>>>>
>>>>>> Thanks!
>>>>>> Pablo
--
Dr. Alessandro Paolini
INFN - CNAF
Viale Berti Pichat 6/2
40127 Bologna
Italy
tel: +39 051 6092723
fax: +39 051 6092916
ICQ: 192172027
skype: alex.paolini
**********************
"credo nel potere del riso e delle lacrime"
"come antidoto all'odio ed al terrore"
"un giorno senza un sorriso"
"è un giorno perso">>> Charlie Chaplin
|