[jira] [Updated] (ACCUMULO-1785) Alter config.sh to optionally just verify environment instead of making changes

["Sean Busbey (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/ACCUMULO-1785?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Sean Busbey updated ACCUMULO-1785:
----------------------------------

    Attachment: ACCUMULO-1785.1.patch.txt

Patch against 1.4.5-SNAPSHOT for changes to implement ACCUMULO_VERIFY_ONLY for bin/config.sh.

Can't get review board to work against the Accumulo project ATM. So I'll just put everything here.

Testing done:

* functional tests
* starting w/o env variable set and no master/slaves correctly started in standalone mode.
* starting w/env variable and no required variables and no master/slaves correctly complained and failed.
* starting w/env variable and required variables and no master/slaves correctly init w/o writing crap, correctly failed to use start-all
* starting with a normal cluster config

One bit I'm not sure about: ${GC} and ${MONITOR}. Right now they're inferred if not present. If we're in VERIFY_ONLY mode, I'm not sure they're actually needed. AFAICT, they're only used in the bin/{start|stop}-* scripts.

Any thoughts about pulling the entire checking/setting of them into the non-verify-only mode?

> Alter config.sh to optionally just verify environment instead of making changes
> -------------------------------------------------------------------------------
>
>                 Key: ACCUMULO-1785
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-1785
>             Project: Accumulo
>          Issue Type: Improvement
>          Components: scripts
>    Affects Versions: 1.4.4, 1.5.0
>            Reporter: Sean Busbey
>            Priority: Minor
>         Attachments: ACCUMULO-1785.1.patch.txt
>
>
> Right now, the accumulo scripts presume that invocation always happens on a node in the cluster. This makes integrating with external cluster management software more difficult.
> The primary issue is that bin/config.sh expects to find a full cluster configuration directory at all times. In the case of a configuration directory that doesn't specify any hosts it takes steps to set up a standalone cluster. That effort is unneeded and undesirable if we're just e.g. trying to go through init from a non-participant host.
> Add an optional environment variable to have config.sh just verify the invariants needed for bin/accumulo to behave correctly. When present, don't make local directories or write to files.



--
This message was sent by Atlassian JIRA
(v6.1#6144)