New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
stacktrace in development environment #31
Comments
Its looks like some issue with etcd configuration. Could you please share the following details:
|
@mkudlej were you able to provide the detail sought by @TimothyAsir ? |
I see the same issue with tendrl-api installed from sources from current master (commit 74db59b). I noticed this during checking of our setup playbooks (see usmqe/usmqe-setup#6 for details) ReproducerI install both etcd and tendrl-api on the same machine, as described in
Then just after the installation (when the playbook listed above finishes), I can't get any answer from the tendrl-api. Both httpd or puma server provides the same answer (so it's not related to httpd configuration):
Further detailsAll details here are provided from the server machine, where both etcd and tendrl-api is isntalled. Restart of tendlr-api doesn't solve the issue:
Configuration of tendrl-api:
The address listed there matches the global ipv4 address of the server machine:
Status of tendrl-apid service:
Then there is etcd, which seems to run fine:
And it's configuration:
Note that etcd configuration for tendrl is not yet properly documented, see:
so that the setup we do right now is based on tendrl demo sessions, see tendrl role, for details. Also note that when I use the same etcd configuration with tendrl-api package, the api works fine. |
@TimothyAsir @sankarshanmukhopadhyay I'm providing the details instead of mkudlej in the previous comment. |
@anivargi I remember even if tendrl api and etcd services are running the api |
Any updates on this? |
@mkudlej , can you respond to #31 (comment) please. There is an issue with tendrl-apid if there is no nodes are present in etcd, it always throws 500. Could you please confirm this? |
@nthomas-redhat Do I read it right that you are asking for us to reproduce this again and check if there are some nodes in etcd? How could I check that there are no nodes in etcd? I'm not familiar with such implementation details yet. That said, there are multiple problem we need to fix. When some etcd key/value is missing, tendrl-api should
|
When I use
I checked that tendrl-node-agent service is up and running on all storage nodes. To verify whether I can reproduce the issue when tendrl-node-agent have no chance to talk to etcd (as asked in #31 (comment)), I try the setup again, but without installing and setting up node-agents (dropping tendrl-node-agent and tendrl-gluster-integration roles from the playbook). Then I see this:
with the same traceback in the logs:
Which means that the answer to @nthomas-redhat 's question is that yes, the issue seems to be caused by node-agent not being up and running (there must have been some other, now fixed, issue which caused node-agent to fail previously). With this in mind, I think that we can agree that:
|
This seems to conflict with the "green field" use case where I want to install ceph or gluster (by depending on a node agent running for an as yet non-existent service). |
@mkudlej do we need to reevaluate this to check if the problem here is still relevant after rescope? |
I've tried to install all components from sources by Ansible playbooks
from https://github.com/Tendrl/usmqe-setup.
I see this stacktrace when I call
The text was updated successfully, but these errors were encountered: