[
https://issues.apache.org/jira/browse/METRON-472?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15531764#comment-15531764
]
Otto Fowler edited comment on METRON-472 at 9/29/16 4:41 AM:
-------------------------------------------------------------
What I'm seeing is that if you run vagrant provision ( as you will be told in
the failure message ) with the same parameters as run.sh passes to vagrant up,
it will or possibly will work.
So I'm doing this tonight as I test out my deployment changes -
>./run.sh
...
...
fail - because the broker isn't started or something and there is +1 partitions
than brokers....
>./runAgain.sh
..
..
TASK [deployment-report : set_fact] ********************************************
ok: [node1]
TASK [deployment-report : debug] ***********************************************
ok: [node1] => {
"Success": [
"Apache Metron deployed successfully",
" Metron @ http://node1:5000";,
" Ambari @ http://node1:8080";,
" Sensor Status @ http://node1:2812";,
" Topology Status @ http://node1:2812";,
" Zookeeper @ node1:2181",
" Kafka @ node1:6667",
"For additional information, see https://metron.incubator.apache.org/'"
]
}
----
I thought to create the script because the first time this happened, I _did_
just do vagrant provision, as I would do with the full dev - and it did
*nothing*, because the tags were not passed and the box was running and up.
Provision is run at the END of up, after the box is created configured and
started to install packages. Calling provision and not up saves you the box
build out.
was (Author: ottobackwards):
What I'm seeing is that if you run vagrant provision ( as you will be told in
the failure message ) with the same parameters as run.sh passes to vagrant up,
it will or possibly will work.
So I'm doing this tonight as I test out my deployment changes -
>./run.sh
...
...
fail - because the broker isn't started or something and there is +1 partitions
than brokers....
>./runAgain.sh
..
..
TASK [deployment-report : set_fact] ********************************************
ok: [node1]
TASK [deployment-report : debug] ***********************************************
ok: [node1] => {
"Success": [
"Apache Metron deployed successfully",
" Metron @ http://node1:5000";,
" Ambari @ http://node1:8080";,
" Sensor Status @ http://node1:2812";,
" Topology Status @ http://node1:2812";,
" Zookeeper @ node1:2181",
" Kafka @ node1:6667",
"For additional information, see https://metron.incubator.apache.org/'"
]
}
----
I thought to create the script because the first time this happened, I _did_
just do vagrant provision, as I would do with the full dev - and it did
*nothing*, because the tags were not passed and the box was running and up.
> Vagrant Quick Dev should have both run.sh and runAgain.sh
> ---------------------------------------------------------
>
> Key: METRON-472
> URL: https://issues.apache.org/jira/browse/METRON-472
> Project: Metron
> Issue Type: Improvement
> Affects Versions: 0.2.1BETA
> Environment: Vagrant Quick Dev
> Reporter: Otto Fowler
>
> The vagrant quick dev is executed by running run.sh ( as opposed to just
> doing vagrant up in the full dev ). run.sh does the vagrant up, with the
> correct tags declared and the skip solr flag.
> Often the vagrant up does not finish execution, and as recommended you have
> to do a vagrant provision to continue.
> Depending on where in the process you have failed however, you will not
> succeed because you have not declared the same tags and skips.
> There should be another script to be run in the provision/retry case.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
