The above-mentioned combination is the one which is configured in the HEAT template at the time of writing of this WIKI page and may cause an error when starting A&AI. Indications might be that you get errors running "demo.sh init" and/or do not see a customer when running VID as user demo and searching for service instances. Same can be that after deploying a service instance, the next page hangs with "Retrieving .... <some uuid>"
There are a couple of steps to fix all this which is described in various places and which I try to assemble on this one page. Some steps might be unnecessary, wrong or stupid. But this is how I got AAI to work.
This is for the one instance AAI. I'll update as soon as the docker repository allows me to try the two-instance aai.
In /opt/aai_vm_init change the docker run commands (I put a comment before the original one in case I want to revert)
#docker run -d --net=host --name="hbase-1.2.3" $NEXUS_DOCKER_REPO/aaidocker/aai-hbase-1.2.3 |
|---|
Observer the docker logs with "docker logs -f aai-service". If you see lines like
updateQueryData.sh: failed for service-design-and-creation/models/model/c5171ae0-44fb-4c04-b482-d56702241a44 /opt/app/aai/bundleconfig/etc/scriptdata/widget-model-json/segmentation-assignment-1.0.json on cmd 19
you might want to look the log files which are listed as
See output and error file: /opt/app/aai/logs/misc/updateQueryData.sh.log.2017-06-14
Use this command to see the contents of the log file:
| docker exec -it aai-service cat /opt/app/aai/logs/misc/updateQueryData.sh.log.2017-06-14 |
|---|
You might see lines like
GET service-design-and-creation/models/model/5e68299a-79f2-4bfb-8fbc-2bae877a2459 returned AAI_7402
details java.net.ConnectException: Connection refused (Connection refused)-Error during GET
This means that the putTool.sh script tried to access the server before it started accepting requests. To avoid this error we need to increase the sleep timeout between the starting of server and execution of putTool.sh.
docker exec -it aai-service bash vi /startup.sh |
|---|
Update the sleep time from 20 to 300 (worked on my machine could differ for your machine).
These steps made my ONAP instance work up to the point where I could add VF to a service instance (which was not possible before)
1 Comment
Nayan Deshmukh
In case of aai-service if I use "release-1.0.0" as the docker_gitbranch then the files are created and the scripts are placed at the right places by the chef cookbooks. The only real problem that I faced was that putTool.sh was being executed even before the server was up, so I just updated the sleep between their execution.
Also in my case I had to modify the docker run command to use the --dns argument to set the dns vm ip for the aai-service docker. We can update the init script if others are seeing similar problems.