[one-users] reg oneflow - service template

Rajendar K k.rajendar at gmail.com
Tue Dec 3 17:40:39 PST 2013


Hi Carlos,
                Thanks for the mail. Here are my queries,


On Tue, Dec 3, 2013 at 7:52 PM, Carlos Martín Sánchez <
cmartin at opennebula.org> wrote:

> Hi,
>
> On Mon, Dec 2, 2013 at 4:51 AM, Rajendar K <k.rajendar at gmail.com> wrote:
>
>> Hi Carlos,
>>                   I have updated the files [oneflow-templates and
>> service] as per your instructions. Hereby specify the output of oneflow
>>
>> => Elasticty conditions
>>
>>  "min_vms": 1,
>>       "max_vms": 3,
>> *      "cooldown": 60,         => At what period , this parameter is
>> being employed?               *
>>       "elasticity_policies": [
>>         {
>>           "type": "CHANGE",
>>           "adjust": 1,
>>           "expression": "CPU < 60",
>>           "period": 3,
>>           "period_number": 30,
>>           "cooldown": 30
>>         }
>>       ],
>>       "scheduled_policies": [
>>
>>       ]
>>     }
>>
>>
>> kindly provide detail on how auto-scaling is happened in the above
>> sample, with relates to "period_number" , "period" and "cooldown"
>>
>
> The period_number, period and cooldown attributes are explained in detail
> in the documentation [1].
> In your example, the expression CPU < 60 must be true 30 times, each 3
> seconds.
>


*So the scaling should be triggered after 1.30 minutes (30 * 3 seconds) (90
seconds)  is it right?*

*The log shows that scaling is triggered after 16 minutes,*

LOG MESSAGES

12/02/13 10:06 [I] New state: DEPLOYING
*12/02/13 10:07 [I] New state: RUNNING*
*12/02/13 10:22 [I] Role role1 scaling up from 1 to 2 nodes*
12/02/13 10:22 [I] New state: SCALING
12/02/13 10:23 [I] New state: COOLDOWN
12/02/13 10:23 [I] New state: RUNNING
*12/02/13 10:38 [I] Role role1 scaling up from 2 to 3 nodes*
12/02/13 10:38 [I] New state: SCALING
12/02/13 10:39 [I] New state: COOLDOWN
12/02/13 10:39 [I] New state: RUNNING



>
> After the scaling, your service will be in the cooldown period for 30
> seconds before returning to running. The only defined policy is overriding
> the default cooldown of 60 that you set.
>
>
if my understanding is correct, if we didn't specify any cooldown for each
role , it takes the default policy as "60" using that parameter?

In a previous email you mentioned:
>
> On Fri, Nov 29, 2013 at 5:19 AM, Rajendar K <k.rajendar at gmail.com> wrote:
>
>> (ii) Scheduled policies
>>          -> For my trial i have specified
>>                          start time -> 2013-11-29 10:30:30
>>                   is my time format is correct? unable to scale up/down
>> using this time format.
>>
>
> But the output you provided does not have any scheduled_policies. Was that
> another different template?
>

Yes, its a  different template


>
> Regards.
>
> [1] http://opennebula.org/documentation:rel4.4:appflow_elasticity
> --
> Carlos Martín, MSc
> Project Engineer
> OpenNebula - Flexible Enterprise Cloud Made Simple
> www.OpenNebula.org <http://www.opennebula.org/> | cmartin at opennebula.org
>  | @OpenNebula <http://twitter.com/opennebula> <cmartin at opennebula.org>
>
>
>
> On Mon, Dec 2, 2013 at 4:51 AM, Rajendar K <k.rajendar at gmail.com> wrote:
>
>> Hi Carlos,
>>                   I have updated the files [oneflow-templates and
>> service] as per your instructions. Hereby specify the output of oneflow
>>
>>
>> root at aaaa:/srv/cloud/one# oneflow show 179
>> SERVICE 179
>> INFORMATION
>> ID                  : 179
>> NAME                : Sampletest
>> USER                : root
>> GROUP               : oneadmin
>> STRATEGY            : straight
>> SERVICE STATE       : RUNNING
>> SHUTDOWN            : shutdown
>>
>> PERMISSIONS
>>
>> OWNER               : um-
>> GROUP               : ---
>> OTHER               : ---
>>
>> ROLE role1
>> ROLE STATE          : RUNNING
>> VM TEMPLATE         : 590
>> CARNIDALITY         : 3
>> MIN VMS             : 1
>> MAX VMS             : 3
>> COOLDOWN            : 60s
>> SHUTDOWN            : shutdown
>> NODES INFORMATION
>>  VM_ID NAME                    STAT UCPU    UMEM
>> HOST                       TIME
>>    568 role1_0_(service_179)   runn    0   1024M 10.1.26.32
>> 0d 01h24
>>    569 role1_1_(service_179)   runn    0   1024M 10.1.26.31
>> 0d 01h07
>>    570 role1_2_(service_179)   runn    0   1024M 10.1.26.32
>> 0d 00h51
>>
>> ELASTICITY RULES
>>
>> ADJUST       EXPRESSION                                       EVALS
>> PERIOD  COOL
>> + 1          CPU[0.0] < 60                                    30 /
>> 3s   30s
>>
>>
>> LOG
>> MESSAGES
>> 12/02/13 10:06 [I] New state: DEPLOYING
>> 12/02/13 10:07 [I] New state: RUNNING
>> 12/02/13 10:22 [I] Role role1 scaling up from 1 to 2 nodes
>> 12/02/13 10:22 [I] New state: SCALING
>> 12/02/13 10:23 [I] New state: COOLDOWN
>> 12/02/13 10:23 [I] New state: RUNNING
>> 12/02/13 10:38 [I] Role role1 scaling up from 2 to 3 nodes
>> 12/02/13 10:38 [I] New state: SCALING
>> 12/02/13 10:39 [I] New state: COOLDOWN
>> 12/02/13 10:39 [I] New state: RUNNING
>>
>> => Elasticty conditions
>>
>>  "min_vms": 1,
>>       "max_vms": 3,
>> *      "cooldown": 60,         => At what period , this parameter is
>> being employed?               *
>>       "elasticity_policies": [
>>         {
>>           "type": "CHANGE",
>>           "adjust": 1,
>>           "expression": "CPU < 60",
>>           "period": 3,
>>           "period_number": 30,
>>           "cooldown": 30
>>         }
>>       ],
>>       "scheduled_policies": [
>>
>>       ]
>>     }
>>
>>
>> kindly provide detail on how auto-scaling is happened in the above
>> sample, with relates to "period_number" , "period" and "cooldown"
>>
>> with regards
>> Raj
>>
>>
>> Raj,
>>
>> Believe Yourself...
>>
>>
>> On Sat, Nov 30, 2013 at 2:19 AM, Carlos Martín Sánchez <
>> cmartin at opennebula.org> wrote:
>>
>>> Hi,
>>>
>>> On Fri, Nov 29, 2013 at 5:19 AM, Rajendar K <k.rajendar at gmail.com>wrote:
>>>
>>>> Hi All,
>>>>                   I am using opennebula 4.2, i have the following
>>>> queries related to auto scaling features
>>>>
>>>> (i) Even when conditions met, scaling up takes long time to trigger.
>>>>         For my trial,  I have used #periods 5   #period 30     cooldown
>>>> 30
>>>>
>>>>         -> i can able to c the values 5/5 in the log,  and then it
>>>> crosses 16/5 in course of time
>>>>         -> Unable to trace down when it trigger actually.
>>>>         hope it should trigger when it reaches 5/5 ?
>>>>
>>>>
>>> This is a known bug, the period and period number inputs in sunstone are
>>> inverted [1].
>>> To solve it, execute:
>>>
>>> $ sudo wget
>>> http://dev.opennebula.org/projects/opennebula/repository/revisions/5c7b370f6be5f4f9167d508bdd31b580f05dafd4/raw/src/sunstone/public/js/plugins/oneflow-templates.js-O /usr/lib/one/sunstone/public/js/plugins/oneflow-templates.js
>>>
>>> $ sudo wget
>>> http://dev.opennebula.org/projects/opennebula/repository/revisions/970c34464a1d5a71f3a9dfdd3d77d62883400857/raw/src/sunstone/public/js/plugins/oneflow-services.js-O /usr/lib/one/sunstone/public/js/plugins/oneflow-services.js
>>>
>>> Or wait a few days and upgrade to OpenNebula 4.4 :)
>>>
>>>
>>>
>>>> (ii) Scheduled policies
>>>>          -> For my trial i have specified
>>>>                          start time -> 2013-11-29 10:30:30
>>>>                   is my time format is correct? unable to scale up/down
>>>> using this time format.
>>>>
>>>>
>>> Could you paste the output of oneflow show -j ?
>>>
>>>
>>>> (ii) is it like "AND" conditions, if i specify both (i) and (ii)?. I
>>>> have tried separately (i) and (ii) but still in-vain.
>>>>
>>>
>>> No, they do not interfere with each other. Each condition (time, metrics
>>> expression) is evaluated independently.
>>>
>>> Regards
>>>
>>> [1] http://opennebula.org/documentation:rel4.2:known_issues
>>> --
>>> Carlos Martín, MSc
>>> Project Engineer
>>> OpenNebula - Flexible Enterprise Cloud Made Simple
>>> www.OpenNebula.org <http://www.opennebula.org/> | cmartin at opennebula.org
>>>  | @OpenNebula <http://twitter.com/opennebula>
>>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opennebula.org/pipermail/users-opennebula.org/attachments/20131204/06af4de4/attachment-0002.htm>


More information about the Users mailing list