(we got the bug hunted in chat in gmail with Nicolas before my
previous email) and it look like all dumps were good :
[<service.Service object at 0xc189244>, <service.Service object at
0xc18b114>, <service.Service object at 0xc1872bc>, <service.Service
object at 0xc189f24>, <service.Service object at 0xc1887c4>,
<service.Service object at 0xc18bb94>, <service.Service object at
0xc185c2c>, <service.Service object at 0xc18a81c>, <service.Service
object at 0xc18584c>]
Impact: webdutux/disque-home
webdutux
Impact: webdutux/Charge IO disques
webdutux
Impact: webdutux/disque-var
webdutux
Impact: webdutux/swap
webdutux
Impact: webdutux/disque-opt
webdutux
Impact: webdutux/Reseau
webdutux
Impact: webdutux/disque-usr
webdutux
Impact: webdutux/Charge machine
webdutux
Impact: webdutux/disque-slash
webdutux
Traceback (most recent call last):
File "./shinken-scheduler.py", line 495, in <module>
p.main()
File "./shinken-scheduler.py", line 412, in main
self.sched.run()
File "/opt/shinken-20100720/src/scheduler.py", line 737, in run
f()
File "/opt/shinken-20100720/src/scheduler.py", line 552, in consume_results
item.consume_result(c)
File "/opt/shinken-20100720/src/schedulingitem.py", line 602, in consume_result
self.no_more_a_problem()
File "/opt/shinken-20100720/src/schedulingitem.py", line 157, in
no_more_a_problem
impact.deregister_a_problem(self)
File "/opt/shinken-20100720/src/schedulingitem.py", line 226, in
deregister_a_problem
self.source_problems.remove(pb)
ValueError: list.remove(x): x not in list
Sent at 2:21 PM on Monday
So it look like goods values in impacts and source_problems. So it
mean that the host in source_rpoblem was not the good one!
For such bugs (the host seems good, but it's not the real one) there
is one direction to look at : retention load. From now, the rention
load with Python cPickle format (easier to code) but it load every
things. And it looks like it can duplicate objects (strange behaviour
for cPicle in fact). So a solution from now is:
in scheduler.py, find the not_loading tab and add 'source_problems',
'impacts' (it's a tab to say "do not load theses properties from
retention).
not_loading = ['act_depend_of', 'chk_depend_of', 'checks_in_progress', \
'downtimes', 'host', 'next_chk',
'act_depend_of_me', \
'chk_depend_of_me', 'services',
'source_problems', 'impacts']
This is a bad conception of such a loading function. I'll add a
parameter in the properties and running_properties tab of Host and
Service class so we tag explicitly wich properties we load, and not
which one we do not load. It's another proof that all properties
properties must be in theses tab, and not dispatched in the code :D
It's will be ok for the next releasen I put this in the top of the
trac, and will be ok this evening, so there will not be anymore "But
my host pointer is not my host pointer" problems :)
And thanks to Nicolas to find this bug (and another one about a bad
format of a string in satellites.py too) :)
Jean
On Mon, Jul 26, 2010 at 2:17 PM, nap <[email protected]> wrote:
> Add this in the no_more_a_problem function :
>
> print "Here are my impacts", self.impacts
> for i in self.impacts:
> print "Impact:", i.get_dbg_name()
> for p in i.source_problems:
> print p.get_dbg_name()
>
>
> Jean
>
>
> On Mon, Jul 26, 2010 at 2:05 PM, nap <[email protected]> wrote:
>> Oh it was to do not send too much mail in the mailing list, but if
>> other want to folow the debuging phase, we can get back onlist, ti's
>> not a problem :)
>>
>> The idea here is to find why we remove twice the problem (host
>> pointer) from an impact. Impacts list is a list of a set (way to have
>> unique element list) so it should not have duplication of impacts in
>> the list. I ask Nicolas to look at this list before we deregister. If
>> it's a true unique list (and itshould be), it can be that the problem
>> is deregister twice (it can be when he goes UP, but also if a problem
>> register for it, so it's now an impact, so t deregister from it's own
>> impacts). But I don't see how it can bug here. But it does, so I must
>> miss something :)
>>
>>
>> Jean
>>
>> On Mon, Jul 26, 2010 at 1:55 PM, Michael Friedrich
>> <[email protected]> wrote:
>>> nap wrote:
>>>> Ok, let's start for debugging. I send you offlist instructions to find it.
>>>>
>>>
>>> Why offlist? Maybe others want to stumble around too :p
>>>
>>> Kind regards,
>>> Michael
>>>
>>>> When the debug will be finised, we will post the bug and tha patch :)
>>>>
>>>>
>>>> Jean
>>>>
>>>> On Mon, Jul 26, 2010 at 1:50 PM, NICOLAS DUPEUX
>>>> <[email protected]> wrote:
>>>>
>>>>> No problem to reproduce, it's just a matter of seconds after loading the
>>>>> configuration.
>>>>>
>>>>> Here is the stack trace without my patch :
>>>>>
>>>>> Traceback (most recent call last):
>>>>> File "./shinken-scheduler.py", line 495, in<module>
>>>>> p.main()
>>>>> File "./shinken-scheduler.py", line 412, in main
>>>>> self.sched.run()
>>>>> File "/opt/shinken-20100720/src/scheduler.py", line 737, in run
>>>>> f()
>>>>> File "/opt/shinken-20100720/src/scheduler.py", line 552, in
>>>>> consume_results
>>>>> item.consume_result(c)
>>>>> File "/opt/shinken-20100720/src/schedulingitem.py", line 597, in
>>>>> consume_result
>>>>> self.no_more_a_problem()
>>>>> File "/opt/shinken-20100720/src/schedulingitem.py", line 152, in
>>>>> no_more_a_problem
>>>>> impact.deregister_a_problem(self)
>>>>> File "/opt/shinken-20100720/src/schedulingitem.py", line 221, in
>>>>> deregister_a_problem
>>>>> self.source_problems.remove(pb)
>>>>> ValueError: list.remove(x): x not in list
>>>>>
>>>>> Do you want others elements ?
>>>>>
>>>>> Regards
>>>>>
>>>>> ----- "nap"<[email protected]> a écrit :
>>>>>
>>>>>
>>>>>> Thanks.
>>>>>>
>>>>>> Is it reproductible? It should not be twice "deregister" of a
>>>>>> problem,
>>>>>> so there is a problem :)
>>>>>>
>>>>>>
>>>>>> Jean
>>>>>>
>>>>>> On Mon, Jul 26, 2010 at 11:20 AM, NICOLAS DUPEUX
>>>>>> <[email protected]> wrote:
>>>>>>
>>>>>>> Hi,
>>>>>>>
>>>>>>> This patch allow my scheduler to live more than a couple of
>>>>>>>
>>>>>> seconds.
>>>>>>
>>>>>>> It shouldn't be apply as is because it doesn't solve the root cause
>>>>>>>
>>>>>> (why the problem isn't in the list)
>>>>>>
>>>>>>> Regards,
>>>>>>>
>>>>>>> --
>>>>>>> Nicolas DUPEUX<[email protected]>
>>>>>>> Arkea - Domaine Systèmes
>>>>>>> tel : 02.98.00.36.68
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>>
>>>>>>> The Palm PDK Hot Apps Program offers developers who use the
>>>>>>> Plug-In Development Kit to bring their C/C++ apps to Palm for a
>>>>>>>
>>>>>> share
>>>>>>
>>>>>>> of $1 Million in cash or HP Products. Visit us here for more
>>>>>>>
>>>>>> details:
>>>>>>
>>>>>>> http://ad.doubleclick.net/clk;226879339;13503038;l?
>>>>>>> http://clk.atdmt.com/CRS/go/247765532/direct/01/
>>>>>>> _______________________________________________
>>>>>>> Shinken-devel mailing list
>>>>>>> [email protected]
>>>>>>> https://lists.sourceforge.net/lists/listinfo/shinken-devel
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> The Palm PDK Hot Apps Program offers developers who use the
>>>>>> Plug-In Development Kit to bring their C/C++ apps to Palm for a share
>>>>>>
>>>>>> of $1 Million in cash or HP Products. Visit us here for more details:
>>>>>> http://ad.doubleclick.net/clk;226879339;13503038;l?
>>>>>> http://clk.atdmt.com/CRS/go/247765532/direct/01/
>>>>>> _______________________________________________
>>>>>> Shinken-devel mailing list
>>>>>> [email protected]
>>>>>> https://lists.sourceforge.net/lists/listinfo/shinken-devel
>>>>>>
>>>>> --
>>>>> Nicolas DUPEUX<[email protected]>
>>>>> Arkea - Domaine Systèmes
>>>>> tel : 02.98.00.36.68
>>>>>
>>>>> --
>>>>> Ce message et toutes les pieces jointes (ci-apres le "message") sont
>>>>> confidentiels et etablis a l'intention exclusive de ses destinataires.
>>>>> Toute utilisation ou diffusion non autorisee est interdite. Tout
>>>>> message etant susceptible d'alteration, l'emetteur decline toute
>>>>> responsabilite au titre de ce message s'il a ete altere, deforme ou
>>>>> falsifie.
>>>>> -----------------------------------
>>>>> This message and any attachments (the "message") are confidential and
>>>>> intended solely for the addressees. Any unauthorised use or
>>>>> dissemination is prohibited. As e-mails are susceptible to alteration,
>>>>> the issuer shall not be liable for the message if altered, changed
>>>>> or falsified.
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> The Palm PDK Hot Apps Program offers developers who use the
>>>>> Plug-In Development Kit to bring their C/C++ apps to Palm for a share
>>>>> of $1 Million in cash or HP Products. Visit us here for more details:
>>>>> http://ad.doubleclick.net/clk;226879339;13503038;l?
>>>>> http://clk.atdmt.com/CRS/go/247765532/direct/01/
>>>>> _______________________________________________
>>>>> Shinken-devel mailing list
>>>>> [email protected]
>>>>> https://lists.sourceforge.net/lists/listinfo/shinken-devel
>>>>>
>>>>>
>>>> ------------------------------------------------------------------------------
>>>> The Palm PDK Hot Apps Program offers developers who use the
>>>> Plug-In Development Kit to bring their C/C++ apps to Palm for a share
>>>> of $1 Million in cash or HP Products. Visit us here for more details:
>>>> http://ad.doubleclick.net/clk;226879339;13503038;l?
>>>> http://clk.atdmt.com/CRS/go/247765532/direct/01/
>>>> _______________________________________________
>>>> Shinken-devel mailing list
>>>> [email protected]
>>>> https://lists.sourceforge.net/lists/listinfo/shinken-devel
>>>>
>>>>
>>>
>>>
>>> --
>>> DI (FH) Michael Friedrich
>>>
>>> Vienna University Computer Center
>>> Universitaetsstrasse 7 A-1010 Vienna, Austria
>>>
>>> email: [email protected]
>>> phone: +43 1 4277 14359
>>> fax: +43 1 4277 14279
>>> web: http://www.univie.ac.at/zid
>>>
>>> Icinga Core& IDOUtils Developer
>>> http://www.icinga.org
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> The Palm PDK Hot Apps Program offers developers who use the
>>> Plug-In Development Kit to bring their C/C++ apps to Palm for a share
>>> of $1 Million in cash or HP Products. Visit us here for more details:
>>> http://ad.doubleclick.net/clk;226879339;13503038;l?
>>> http://clk.atdmt.com/CRS/go/247765532/direct/01/
>>> _______________________________________________
>>> Shinken-devel mailing list
>>> [email protected]
>>> https://lists.sourceforge.net/lists/listinfo/shinken-devel
>>>
>>
>
------------------------------------------------------------------------------
The Palm PDK Hot Apps Program offers developers who use the
Plug-In Development Kit to bring their C/C++ apps to Palm for a share
of $1 Million in cash or HP Products. Visit us here for more details:
http://ad.doubleclick.net/clk;226879339;13503038;l?
http://clk.atdmt.com/CRS/go/247765532/direct/01/
_______________________________________________
Shinken-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/shinken-devel
