Re: [netvirt-dev] Random errors while fetching JOLOKIA URLs in 3 node CSIT

Tom Pantelis

On Thu, Jul 5, 2018 at 6:29 PM, Jamo Luhrsen <jluhrsen@...> wrote:
Here is a jira to track this one:

This is really a AAA issue so moved to


On 07/03/2018 07:16 PM, Tom Pantelis wrote:

On Tue, Jul 3, 2018 at 9:39 PM, Vishal Thapar <vthapar@... <mailto:vthapar@...>> wrote:

    On Wed, Jul 4, 2018 at 1:01 AM, Michael Vorburger <vorburger@... <mailto:vorburger@...>> wrote:

        On Tue, Jul 3, 2018 at 7:21 AM, Faseela K <faseela.k@... <mailto:faseela.k@...>>wrote:


                I couldn’t follow your response below completely.____

                 But, the URL whatever I am using in CSIT works most of the times, and only in some random cases it
            throws the shiro error.____

                When I updated the CSIT to retry on failure, it works on subsequent retries.____

                 Below is the exception stack trace :

        I don't know the details of what this CSIT wants from
        but I don't suppose it could (now) be replaced with GET /diagstatus from <> ?  That
        requires no authentication, so it would never suffer from this particular problem. Just a thought.

    ​It gives information about which node is leader and which is follower for a given shard while diagstatus only gives
    information about if all services are up or not. It is used in monitoring tool of ODLTools.

    Unless every bit of information we get from jolokia can be provided through an alternate rest call, this needs to be


The exception emanates from shiro code so that would be the place to start. Anyone have cycles now to try to look into this?

aaa-dev mailing list

Join to automatically receive all group messages.