TECHZEN Zenoss User Community ARCHIVE  

Zenoss CE 1.6.3/6.3.2 r317 - Zenoss Core Web UI - CC tab - HTTPError - HTTP Erro ...

Subject: Zenoss CE 1.6.3/6.3.2 r317 - Zenoss Core Web UI - CC tab - HTTPError - HTTP Error 502
Author: Sam Urai
Posted: 2021-05-17 00:33

Hi,

This is a new deployment on CE version.  Everything in CC  UI is good.  In the Zenoss Core Web UI  accessible via the default zenoss5 public endpoint. When I click on the "Control Center" tab, I see  HTTPError - HTTP Error 502 - Bad Gateway followed by "Error fetching daemon list: Bad Gateway"

What could be the cause of it?


Thanks

------------------------------
Sam
------------------------------


Subject: RE: Zenoss CE 1.6.3/6.3.2 r317 - Zenoss Core Web UI - CC tab - HTTPError - HTTP Error 502
Author: Michael Rogers
Posted: 2021-05-17 11:16

Sam,

Is this a single host install (Zenoss Core installed on a single Control Center master) or a multi-host install (delegate hosts added to a Resource Pool, Zenoss Core installed to that pool)?

In either case, check the contents of the /etc/default/serviced config file for the SERVICED_ENDPOINT line.  It should list the IP address of the Control Center master host.  If it lists another host, or if it points to the CC master by name instead of IP, change it to point to the master IP, then do the following:

Shut down all Zenoss services (serviced service stop zenoss)
Shut down Control Center on all hosts (systemctl stop serviced)
Confirm no Docker containers are running (docker ps -qa)
Shutdown any stale Docker containers (docker rm $(docker ps -qa))
Start Control Center (systemctl start serviced)
Start Zenoss services (serviced service start zenoss)

Let me know your results?


------------------------------
Michael J. Rogers
Senior Instructor - Zenoss
Austin TX
------------------------------


Subject: RE: Zenoss CE 1.6.3/6.3.2 r317 - Zenoss Core Web UI - CC tab - HTTPError - HTTP Error 502
Author: Sam Urai
Posted: 2021-05-17 13:56

I had updated my first post but probably got deleted I guess. Here is the container log.  This is a multi-host installation.  A 'master' pool with one CC master.  A ZenossCore pool with 3 delegates.  Zookeeper ensemble  - CC master + 2 delegates ( 2 of the 3 delegates from the ZC pool ).  For the  'SERVICED_ENDPOINT" variable. I've the FQDN of the CC master. All these hosts are hosted on the same one hypervisor.  Thanks

E0517 17:47:05.188092 00001 proxy.go:250] Container address is invalid. Can't create proxy: master.cc.com:443
2021/05/17 17:47:10 200 2.114707ms POST /api/metrics/store
2021/05/17 17:47:24 http: proxy error: read tcp 127.0.0.1:47348->127.0.0.1:5443: read: connection reset by peer
2021/05/17 17:47:25 200 1.898842ms POST /api/metrics/store
E0517 17:47:26.507626 00001 proxy.go:250] Container address is invalid. Can't create proxy: master.cc.com:443
E0517 17:47:27.278291 00001 proxy.go:250] Container address is invalid. Can't create proxy: master.cc.com:443
​


------------------------------
Sam
------------------------------


Subject: RE: Zenoss CE 1.6.3/6.3.2 r317 - Zenoss Core Web UI - CC tab - HTTPError - HTTP Error 502
Author: Michael Rogers
Posted: 2021-05-17 14:23

Sam,

If you're using the CC master FQDN for SERVICED_ENDPOINT, please change it to the CC master IP address and follow the shutdown/startup steps in my last post.

When complete, please let me know your results?


------------------------------
Michael J. Rogers
Senior Instructor - Zenoss
Austin TX
------------------------------


Subject: RE: Zenoss CE 1.6.3/6.3.2 r317 - Zenoss Core Web UI - CC tab - HTTPError - HTTP Error 502
Author: Sam Urai
Posted: 2021-05-17 14:50

That change did it!    Thanks @Michael.    So the question is - why you can not use FQDN for the endpoint?    The documentation says one can use the hostname or IP.

Please advise. 
Thanks



------------------------------
Sam
------------------------------
E0517 17:47:05.188092 00001 proxy.go:250] Container address is invalid. Can't create proxy: master.cc.com:443
2021/05/17 17:47:10 200 2.114707ms POST /api/metrics/store
2021/05/17 17:47:24 http: proxy error: read tcp 127.0.0.1:47348->127.0.0.1:5443: read: connection reset by peer
2021/05/17 17:47:25 200 1.898842ms POST /api/metrics/store
E0517 17:47:26.507626 00001 proxy.go:250] Container address is invalid. Can't create proxy: master.cc.com:443
E0517 17:47:27.278291 00001 proxy.go:250] Container address is invalid. Can't create proxy: master.cc.com:443
​


------------------------------
Sam


Subject: RE: Zenoss CE 1.6.3/6.3.2 r317 - Zenoss Core Web UI - CC tab - HTTPError - HTTP Error 502
Author: Michael Rogers
Posted: 2021-05-17 15:34

Sam,

You should be able to use the FQDN or hostname but a failure to resolve the name will result in the same message logged: 

Container address is invalid. Can't create proxy: host.domain.tld:443​

If we stick to using strictly IPs we can remove name resolution from the equation.

------------------------------
Michael J. Rogers
Senior Instructor - Zenoss
Austin TX
------------------------------
E0517 17:47:05.188092 00001 proxy.go:250] Container address is invalid. Can't create proxy: master.cc.com:443
2021/05/17 17:47:10 200 2.114707ms POST /api/metrics/store
2021/05/17 17:47:24 http: proxy error: read tcp 127.0.0.1:47348->127.0.0.1:5443: read: connection reset by peer
2021/05/17 17:47:25 200 1.898842ms POST /api/metrics/store
E0517 17:47:26.507626 00001 proxy.go:250] Container address is invalid. Can't create proxy: master.cc.com:443
E0517 17:47:27.278291 00001 proxy.go:250] Container address is invalid. Can't create proxy: master.cc.com:443
​


------------------------------
Sam


Subject: RE: Zenoss CE 1.6.3/6.3.2 r317 - Zenoss Core Web UI - CC tab - HTTPError - HTTP Error 502
Author: Sam Urai
Posted: 2021-05-18 13:07

Yeah you are right. However I've to figure out where that name resolution is failing. In the deployment, I did define our DNS servers with

SERVICED_DOCKER_DNS variable on all CC hosts. 

Thanks



------------------------------
Sam
------------------------------


< Previous
Zenoss CE - High Availability Deployment
  Next
New deployment after crash of old deployment
>