About the Mesos & DC/OS category

Questions & discussions about how to run linkerd inside Mesos and DC/OS.

Well, if you’re already asking…

I’ve deployed Linekrd 1.3.0 on my DC/OS 1.9 environment. “In front” of our external Linekrd we have AWS ELB that configured to send traffic exactly to the agents running Linkerd tasks. I see the tasks running, but when trying to reach the new endpoint (via the old Linkerd instances, still routing traffic for this cluster), I get a clear text response:

java.net.ConnectException: Connection refused: /<IP_ADDRESS>:7770. Remote Info: Upstream Address: /<IP_ADDRESS>:46038, Upstream Client Id: Not Available, Downstream Address: /<IP_ADDRESS>:7770, Downstream Client Id: #/io.l5d.marathon/linkerd-upgrade, Trace Id: f4f49181b3ad83c6.f4f49181b3ad83c6<:f4f49181b3ad83c6

I should say I’ve changed the admin port to be deployed on 7770, not default 9990.

I’m currently checking routing problems between Amazon and our cluster.
The dtab config is loaded as far I can tell and is quite basic.

I managed to reach the delegator. The problem was a wrong config, since I used the current Linkerd-0.9.0 config we have deployed and maybe there were changes made 4 version later. In the admin block I had to explicitly enter the routable 0.0.0.0 address, since the admin endpoint was deployed on internal 127.0.0.1 by default.

hi @jacobgo! sorry we missed this issue, since it was a comment and not filed as a question it doesn’t show up in “Latest” when we check these forums.

ah yeah, I believe this change happened in https://github.com/linkerd/linkerd/pull/1366. Glad you fixed the issue! You can find more about the admin configs here: https://linkerd.io/config/1.3.0/linkerd/index.html#administrative-interface

To file a question about Mesos & DCOS, go to https://discourse.linkerd.io/ and click on “New Topic”. Change the value of the Category dropdown from “Uncategorized” to “Mesos & DC/OS”