Compose file #2
Comments
the env files contain variables specifying (real) addresses to services - I'll update the readme describing what needs to be in those variables
you need to set node labels such that they match the ones in the compose file under |
|
I continued without the env files it eventually also worked when i removed the label constraints from the docker compose file, since i'm only running 1 node to test anyway. Also during the initial build process one of the urls used to get software seems compromised and looks to be a spam-redirect site now. In /docker/sep-wso2is file There doesn't happen to be an .OVA image to deploy on VMware ESXI host as an easier solution? replicas image WSO2 has ALFRESCO has WILDFLY has : sep_wildfly.1.kbn59p9pdn5u@labotest | 2019-05-10 11:31:41,054Z ERROR [org.jboss.as.controller.management-operation] (Controller Boot Thread) {u=, rq=} WFLYCTL0013: Operation ("add") failed - address: ([("subsystem" => "webservices")]) - failure description: {"WFLYCTL0080: Failed services" => {"jboss.ws.config" => "org.jboss.msc.service.StartException in service jboss.ws.config: java.security.ProviderException: Could not initialize NSS KEYCLOAK has sep_keycloak.1.xobb5zx0wo3x@labotest | Setting ownership (dockeru:dockeru) on volume and service dirs |
* add necessary information for deploy preparations to readme * add service dependency variables in compose file and dependency wait script * fix wso2is image to not use internal addresses for resource download github: #2
|
I've pushed a few changes to master as well as a new tag
See Evironment variables section of the readme for explanation of the required variables. I've removed the references to the env files.
See Service placement section of the readme. Removing the constraints on a single node as you did also works.
It's an internal site which you cannot resolve and have forgotten to fix - you should be able to build it now.
No, we currently don't have that.
This was probably due to the fact that some services have dependencies to other services. Please see the Service dependencies section of the readme. You should probably rebuild at least the wso images and delete the volumes, before trying again. |
|
Thanks for all those updates! I started from zero and rebuilt everything, wiped all volumes/images from docker as well. The same servers are still having issues but their errors are now a lot clearer and all of them are waiting /relying on idp that idp container seems to be able to connect to the database and ldap but fails at creating the databases afterwards. sep_idp.1.219fsa9v9fbb@labotest | db:5432 is successfully reached after [1] tries |
|
turns out the jdbc driver for postgres is missing from the image ...working on fix |
jdbc driver was missing from wso2is base image github: #2
|
ok, I've added the missing postgres driver in the base wso image - you'll need to re-build the wso images from the latest master c46833b |
|
Thanks , the wso server works now. However the problem now shifted to alfresco. I've tried to debug, but most i've found is that possible a .jar is included twice and perhaps a different version sep_alfresco.1.e2urmyol5h2c@labotest | SEVERE: Exception sending context initialized event to listener instance of class org.apache.myfaces.webapp.StartupServletContextListener sep_alfresco.1.e2urmyol5h2c@labotest | |
Excude org.slf4j:jcl-over-slf4j:1.7.5 by adding it as a provided dependency in the parent pom. Having it causes class loging issues. github: #2
|
ok, the problem now was with Please try re-building alfresco and try again. |
|
I'm afraid the same error remains. since it was only a change in the pom.xml , i did the change manually, rebuilt alfresco , redeployed. |
|
did you rebuild the docker image only or the java project also? ...they both need to be rebuilt. |
|
ah yes , i only rebuilt de alfresco docker. I'll rebuilt the entire project tomorrow. It does take a bit of time ;) |
|
Thanks again for all your patience ... alfresco working now , next wildfly and UI wait on keycloak. sep_keycloak.1.jth57o039bku@labotest | usermod: no changes |
cli configuration scripts were missing for keycloak github: #2
|
the missing files are there now - please rebuild the keycloak images. |
|
keycloak up and running. wildfly this time |
|
looks like a bug in the alpine openjdk package https://bugs.alpinelinux.org/issues/10126 |
|
I changed in both directories the Dockerfile with : .... RUN rebuild wildfly docker and left the docker swarm , pruned all volumes and redeployed the entire stack But still getting sep_wildfly.1.ntcis9y5osaj@labotest | 2019-05-15 13:07:39,917Z WARN [org.jboss.ws.cxf] (MSC service thread 1-1) {u=, rq=} JBWS024087: Could not early initialize security engine |
due to a bug in alpine package for openjdk, nss needs to be installed manually. https://bugs.alpinelinux.org/issues/10126 github: #2
|
I was able to reproduce the initial error (missing nss lib It looks like you aren't getting the file not found exception anymore though, so probably your fix also worked, but still could you try with the latest master. Also could you provide the full exception, note that |
|
Here's the log through running those commands ################# env ################# Setting ownership (dockeru:dockeru) on volume and service dirs
|
|
Ah, so I guess you still have the same problem (the lib not being installed). |
|
This was from my fix, i'll pull in the entire master branch again maybe and ./build.sh again. |
|
Most errors now are about ERROR: relation "seip_configurations" does not exist sep_wildfly.1.xvshfmovyy3w@labotest | |
for this you'll need to update the environment section of the wildfly service in the compose file add:
If I remember correctly there was some kind of race condition (that I don't think we were able to exactly pin down) which was causing this error. Could you try restarting the service/redeploying the stack. Or it might be because of the previous error. |
|
I've put the new logs in a pastebin, because the thread is getting quite long : It got to DEPLOYMENT COMPLETED! INITIALIZING PLATFORM... |
|
it seem you don't have the key stores configured. |
|
I have One with the certificate and one with the private key. |
|
so I found a default wso2carbon.key file which i imported and wildfly and UI are now also running all services are now running when i go to http://ipoftheserver , i get NGINX 404 error not found ID NAME MODE REPLICAS IMAGE PORTS |
|
could you go into the proxy container and give the nginx configuration under |
This is a default site configuration which will simply return 404, preventingchance access to any other virtualhost.server { } |
|
for some reason the proxy configuration did not get generated. |
|
sep_proxy.1.jynvi8n49ucb@labotest | _=/usr/bin/env |
most likely this is the problem - this binary is responsible for generating the nginx configuration or you could try the commands in the sub-shell manually (now when I think about it it will not fail, if an error occurs, because of the sub-shell) |
|
it fails from cd "$GOPATH/src/git.sirmaplatform.com/stash/seip/http-proxy" I printed the env GOPATH=/root/go |
|
are you running these commands inside the container? |
|
Hi, |
I've tried to run the docker command with the compose file but I get an error about 4 referenced .env files that aren't found.
iiif.env: no such file or directory
keycloak.env: no such file or directory
wildfly.env: no such file or directory
proxy.env: no such file or directory
I've tried to compose with those references removed. But I end up with a swarm with all services in pending mode with the error : no suitable node (scheduling constraints not satisfied on 1 node)
and replicas remain on 0/1
The text was updated successfully, but these errors were encountered: