[ClusterLabs] Antw: Re: Antw: [EXT] Hypothetical question: transitions slow in one direction

[john tillman]

>>>> "john tillman" <johnt at panix.com> schrieb am 14.02.2022 um 15:54 in
> Nachricht
> <7db29102efdf0c227c42bb77c1c95829.squirrel at mail.panix.com>:
>>>  "DC" is the thing that crm_mon displays as "Current DC:" ;‑)
>>
>>
>> LOL ... wow, it *IS* a Monday morning.  Thank you.
>>
>> So I booted my VMs for the experiment and created the cluster as before
>> and now the problem is gone?  The transition is equally fast in both
>> directions.  I repeated the experiment again with the DC on the other
>> node
>> and the issue is gone.
>
> My initial suspect was that when the DC goes stanbdby the election for a
> new
> DC could take extra time...


Thank you.  If it happens again I will look closer at the DC.


>
>>
>> Sorry to have wasted everyone's time.  If email could blush ...
>>
>> Can you take a guess at why a slow down in one direction might have
>> happened?
>>
>>
>>>
>>>
>>>>>> On 02/14/2022 at 02:19 PM, in message <3EEBF8E6.B43 : 250 : 31555>,
>>>>>> "john
>>> tillman" <johnt at panix.com> <johnt at panix.com> wrote:
>>>>>>> "john tillman" <johnt at panix.com> schrieb am 11.02.2022 um 19:01 in
>>>> Nachricht
>>>> <72bdd45d23c0ac29fb3d8751d89b32ee.squirrel at mail.panix.com>:
>>>>> Hypothetically, if I have a two node cluster on identical VMs without
>>>>> fencing and I see transitions take longer in one direction than the
>>>>> other.
>>>>
>>>> How long is "longer", and where is your DC?
>>>
>>>
>>> Less than 10 seconds for the short transition and maybe a minute for
>>> the
>>> longer.
>>>
>>> "DC" ... ok, it's Monday before my first cup of coffee. I can't figure
>>> out what DC means. Sorry, but can you spell it out?
>>>
>>>
>>>>
>>>>> What are some possible reasons?
>>>>>
>>>>> For example, "pcs node standby nodeX" or "pcs cluster stop nodeX"
>>>>> results
>>>>> in a very fast transition to nodeY. However, given the same commands
>>>>> but
>>>>> going from nodeY to nodeX takes much, much longer.
>>>>>
>>>>> And note that powering off either VM does *not* exhibit this behavior
>>>>> and
>>>>> transitions times in both direction are the same, fast.
>>>>>
>>>>> Best regards,
>>>>> â€â€̃John
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Manage your subscription:
>>>>> https://lists.clusterlabs.org/mailman/listinfo/users
>>>>>
>>>>> ClusterLabs home: https://www.clusterlabs.org/
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Manage your subscription:
>>>> https://lists.clusterlabs.org/mailman/listinfo/users
>>>>
>>>> ClusterLabs home: https://www.clusterlabs.org/
>>>>
>>>
>>>
>>> _______________________________________________
>>> Manage your subscription:
>>> https://lists.clusterlabs.org/mailman/listinfo/users
>>>
>>> ClusterLabs home: https://www.clusterlabs.org/
>>>
>>> _______________________________________________
>>> Manage your subscription:
>>> https://lists.clusterlabs.org/mailman/listinfo/users
>>>
>>> ClusterLabs home: https://www.clusterlabs.org/
>>>
>>
>>
>> _______________________________________________
>> Manage your subscription:
>> https://lists.clusterlabs.org/mailman/listinfo/users
>>
>> ClusterLabs home: https://www.clusterlabs.org/
>
>
>
> _______________________________________________
> Manage your subscription:
> https://lists.clusterlabs.org/mailman/listinfo/users
>
> ClusterLabs home: https://www.clusterlabs.org/
>