topic Re: Centos 8 - 502 Bad Gateway in Alfresco Forum

Centos 8 - 502 Bad Gateway

wolfgtweb — Fri, 10 Apr 2020 14:52:21 GMT

I loaded Alfresco on a Centos 8 server yesterday using the docker-compose method and after the install I could not get to the web page. When I try http://localhost:8080/share I get a 502 Bad Gateway message back. In the log, I see this entry:

proxy_1 | 2020/04/10 14:48:18 [error] 7#7: *22 connect() failed (113: Host is unreachable) while connecting to upstream, client: 172.18.0.1, server: , request: "GET /share/ HTTP/1.1", upstream: "http://172.18.0.11:8080/share/", host: "localhost:8080"
proxy_1 | 172.18.0.1 - - [10/Apr/2020:14:48:18 +0000] "GET /share/ HTTP/1.1" 502 157 "-" "Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101 Firefox/68.0"

I thought that I must have done something wrong with the server prior to adding alfresco. So I created an entirely new Centos 8 server installation. Ran updates on it, installed docker and ran the alfresco docker-compose and got the same result. What is going on?

Re: Centos 8 - 502 Bad Gateway

wolfgtweb — Fri, 10 Apr 2020 15:40:40 GMT

Something I didn't notice before. In the live log for Alfresco, every couple of seconds I get this:

2020-04-10 15:35:21.044 ERROR (org.alfresco.solr.AlfrescoCoreAdminHandler@5a865416_Worker-11) [ ] o.a.s.t.AbstractTracker Tracking failed for AclTracker - alfresco
solr6_1 | java.net.NoRouteToHostException: No route to host (Host unreachable)

Followed by a bunch of lines of information. I am not familiar with what this log is supposed to look like, so I was assuming this was normal. But I now see that this is most likely the cause. But why is it doing this on a fresh install? Two fresh installs?

Re: Centos 8 - 502 Bad Gateway

wolfgtweb — Fri, 10 Apr 2020 19:37:37 GMT

Well, I figured it out. It will not work on CentOS 8. But runs perfectly fine on 7.

Re: Centos 8 - 502 Bad Gateway

adamparker — Mon, 27 Apr 2020 05:00:52 GMT

The Docker images are just not working on Centos 8.

On default settings the Share Container can't talk to Postgres and you get SEVERE failure errors.

SOLR goes round in circles with constant errors about NoRouteToHost.

The only way I could get it to work at all, was to start 'docker-compose up', and then immediately disable firewalld. You can't disable firewalld entirely as it needs it to send some iptables commands and fails. The Proxy container also relies on it. So you end up with this ridiculous workaround.