Hi Bharat,

First of all: thanks for reporting the issue!

Would it be possible to write a Marvin test for it, so we can test this 
automatically and prevent regression? Or did you find this by an existing 
Marvin test that I missed?

Please downgrade to critical. We should fix this, and we will, but it shouldn’t 
block 4.6.0 IMHO. Talked to Rajani and she agrees.

Regards,
Remi



On 05/11/15 10:29, "Bharat Kumar" <[email protected]> wrote:

>Hi Wilder,
>
>If we think that it is ok to have the work around for now fix this later, we 
>can bring the priority down to critical.
>
>Thanks,
>Bharat.
>
>On 05-Nov-2015, at 2:49 pm, Wilder Rodrigues <[email protected]> 
>wrote:
>
>> Hi Bharat,
>> 
>> Please check if what you just mention will work. If it does, then the issue 
>> is not a blocker.
>> 
>> Cheers,
>> Wilder
>> 
>> 
>>> On 05 Nov 2015, at 10:09, Bharat Kumar <[email protected]> wrote:
>>> 
>>> I have’t checked, but form the code it looks like it will work if we 
>>> reissue the reset command after the slave becomes master.
>>> 
>>> —Bharat.
>>> 
>>> On 05-Nov-2015, at 12:59 pm, Erik Weber 
>>> <[email protected]<mailto:[email protected]>> wrote:
>>> 
>>> On Thu, Nov 5, 2015 at 7:15 AM, Bharat Kumar 
>>> <[email protected]<mailto:[email protected]>>
>>> wrote:
>>> 
>>> Hi,
>>> 
>>> In my local setup i found this issue, The user VM password is not getting
>>> saved in the backup router.
>>> 
>>> we send the save password command to both the VRs in a rvr enabled
>>> network, But the password gets saved only in the master VR. This happens
>>> because the password server is not running in the backup.
>>> Because of this if someone resets the password of a VM and starts it when
>>> the backup becomes master. Then the password of the user VM will not
>>> change, because the save password command was not successful.
>>> 
>>> This breaks the password reset functionality, I have a raised a Blocker
>>> issue CLOUDSTACK-9035<
>>> https://issues.apache.org/jira/browse/CLOUDSTACK-9035> to track this.
>>> 
>>> 
>>> 
>>> Does it work if you re-issue the password reset call after the slave vr has
>>> been promoted to master?
>>> If it does, then atleast you have a workaround for the issue.
>>> 
>>> --
>>> Erik
>>> 
>> 
>

Reply via email to