[
https://issues.apache.org/jira/browse/ATLAS-2315?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16295093#comment-16295093
]
David Radley commented on ATLAS-2315:
-------------------------------------
Thanks for the pointer [~bpgergo]. This does look similar - I wanted to raise
the wider issues in this Jira and agree potential approaches to resolve
> Atlas says it is successfully started even if the index engine is not
> working.
> -------------------------------------------------------------------------------
>
> Key: ATLAS-2315
> URL: https://issues.apache.org/jira/browse/ATLAS-2315
> Project: Atlas
> Issue Type: Bug
> Environment: Mac dev master
> Reporter: David Radley
>
> Atlas says it is successfully started even if the index engine is not
> working.
> Atlas does not check that that the index engine is working before saying that
> Atlas has successfully started.
> One way to handle this would be to have stages of initialization, including:
> - If embedded Hbase
> - embedded backing store started
> - embedded backing store accepting requests
> - If embedded index
> - embedded index started
> - embedded index accepting requests
> if not embedded index
> - index engine verified to be active.
> - servlet engine started
> - servlet engine accepting Rest calls
> Another approach would be to check that JanusGraph is sane prior to starting
> Atlas and separating out any starting of embedded and verification of Janus
> from any Atlas specific code.
>
> There needs to appropriate message so there is an obvious action to be taken
> when the dependent components are not there when starting atlas using
> atlas_start.py.
> I am currently seeing issues when I have an Atlas built and running with
> embedded Solr and Hbase and Janus. I then re build - the index / Hbase are
> not cleaned up - so I get errors when attempting to start atlas again, all
> the best, David.
>
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
