Issue related to installing OEC on Local Servers

[Anurag Pandey]

Hi Guys,

I was trying to install OEC on my local servers (CentOS 6.10).  I ran the following command (suggested by Installation page):

sudo rpm -i opsgenie-oec-1.1.1.rhel6.x86_64.rpm

and got the below error:

Starting Opsgenie Edge Connector (OEC)INFO[2021-02-10T13:54:24.650075803-05:00] OEC version is 1.1.2
INFO[2021-02-10T13:54:24.650302874-05:00] OEC commit version is ce55d5202eb29611616e32a6921c89e464d2de76
INFO[2021-02-10T13:54:24.650658946-05:00] Visibility timeout cannot be lesser than 15 seconds or greater than 12 hours, default value[30 s.] is set.
INFO[2021-02-10T13:54:24.650797902-05:00] OEC-metrics serves in http://localhost:7070/metrics .
ERROR[2021-02-10T13:54:24.651091827-05:00] OEC-metrics error: listen tcp :7070: bind: address already in use
INFO[2021-02-10T13:54:24.651341524-05:00] Queue processor is starting.
ERROR[2021-02-10T13:54:24.851069741-05:00] Queue processor could not get initial token and will terminate.
FATAL[2021-02-10T13:54:24.851132448-05:00] Token could not be received from Opsgenie, status: 422 Unprocessable Entity, message: {"message":"Key format is not valid!","took":0.0,"requestId":"41dfdd23-c220-4e59-b82e-c18395d91c91"}
warning: %post(opsgenie-oec-rhel6-1.1.1-1.x86_64) scriptlet failed, exit status 1

 The Port 7070 is used by internal applications and that cannot be changed.

Please suggest what could be the solution.

Thanks in advance

[Anurag Pandey · Accepted solution]

The Answer to the above query is to edit the OEC Service file:

ExecStart=/usr/local/bin/OpsgenieEdgeConnector --metric-port 7071 (Any port which you want to use)

[Ben Bieler]

Any updates here? In my case port 7070 is not in use. Do I need to open it in the EC2 security groups?

[Darryl Lee]

Hi Ben,

This is Darryl, Opsgenie Support, I am here to help. :)

If your port 7070 is not in use, the OEC service should be running without any problem.

The other thing required is your OEC can reach out to AWS SQS to poll the tasks, the domain name is varied depends on your Opsgenie hosted region.

Also ensure you have followed the instructions on these two docs to configure your config.json and run it.

https://docs.opsgenie.com/docs/oec-configuration

https://docs.opsgenie.com/docs/oec-running

Hope this helps.

Kind regards,
Darryl Lee
Opsgenie Support

[Luca Cristian Stefan]

I do have the same problem, waiting for atlassian for a solution here.

if you do

netstat -tulpn | grep -i "7070"

It will show you a process of Edge Connector already running but the process of starting it up, cranks down.

I'd be glad to hear an answer for this cause you can't alter the port that app is running on linux as i know.

################## EDIT ###################

I found that Processes from /usr/local/bin/OpsgenieEdgeConnector were already running and keeping the port up (if you kill 7070 is not enough to restart oec)

You have to kill previous failed processes to be able to start it again

If you do

watch -n 1 journalctl -xe

You will see that EVEN if OEC process is killed, worker is alive and working, i think this is a bug.

[connor mcguane]

If port 7070 is already used, you can alter the port the OEC will use in the oecService.json.  At least, that is where that setting is in the Windows install.  I am not sure for the Linux one.