It certainly is supported and is expected to work. Note that it can take upto 2
x leader election timeout (not sure what it's set to, can be seconds to tens of
seconds) for CDS to stabilized after such an excercise.
>> What is the default leader election timeout? Where do we check that?
Currently we are waiting till DATASTORE sync_status becomes true.
But will that fluctuate?
Thanks,
Faseela
-----Original Message-----
From: Robert Varga <[email protected]>
Sent: Wednesday, March 13, 2019 2:49 PM
To: Faseela K <[email protected]>; Jaya Priyadarshini
<[email protected]>; Dayavanti Gopal Kamath
<[email protected]>
Cc: [email protected]; [email protected]
Subject: Re: Netvirt 3 node CSIT HA Tests - failure debugging
On 13/03/2019 08:38, Faseela K wrote:
> +Robert
>
> Robert,
Hello,
>
> Jaya is trying the below steps in a 3 node ODL Cluster :
>
> 1. Take Down ODL1 and ODL2
> 1. Check all VM connectivities are intact even if 2 nodes are
> down 2. Bring Up ODL1 and ODL2, sequentially as below
>
> a. ClusterManagement.Start Single Member /1 msg=up: ODL3, down:
> ODL1, ODL2 wait_for_sync=False/
>
> b. ClusterManagement.Start Single Member /2 msg=up: ODL1, ODL3,
> down: ODL2/
>
> 3. Take Down ODL2 and ODL3, sequentially
>
> Once step 3 is performed, many of the subsequent tests fail, before we
> dig into the details of the specific test cases that fail, wanted to
> know whether the above procedure of
I am not sure what you mean. With two out of three nodes out the cluster cannot
make any progress...
> Bringing UP and Bringing Down ODL nodes are supported, and expected to
> work, or do we have to add any additional delay?
It certainly is supported and is expected to work. Note that it can take upto 2
x leader election timeout (not sure what it's set to, can be seconds to tens of
seconds) for CDS to stabilized after such an excercise.
Regards,
Robert
_______________________________________________
controller-dev mailing list
[email protected]
https://lists.opendaylight.org/mailman/listinfo/controller-dev
