[
https://issues.apache.org/jira/browse/BIGTOP-1175?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13988249#comment-13988249
]
Sean Mackrory commented on BIGTOP-1175:
---------------------------------------
{quote}1. Sorry for my ignorance on this one, but how does the init_function in
accumulo-master.svc get invoked? Is it just via passing 'init' to the
accumulo-master script?{quote}
accumulo-master.svc is run through an init script generator. After installing
accumulo, a user should run 'service accumulo-master init' to do the
initialiation. They can then star the master with 'service accumulo-master
start', and use similar commands for stopping, status-checking, etc... the
master and all other services.
{quote}
2. Using /user/accumulo for the "data directory" in HDFS is better than
/accumulo IMO. It also alleviates some of the extra chmod'ing in the
aformentioned init_function.
3. Bad default value in accumulo-env.sh for HADOOP_PREFIX ("/usrlib/hadoop")
4. Remove the "$ACCUMULO_HOME/src/.." lines from general.classpaths in
accumulo-site.xml.
5. In generic_logger.xml, update the value for the "port" param in the
SocketAppender definition from "4560" to "$
{org.apache.accumulo.core.host.log.port}
". This is internally advertised in >=1.5.1
6. In install_accumulo.sh, ACCUMULO_LOG_HOST can be removed for same reason as
#5
{quote}
Good ones - I'll address all these points in a future revision of that patch.
{quote}7. Is there a reason that the accumulo (unix) user's home directory is
set to /var/lib/accumulo instead of /usr/lib/accumulo?{quote}
Sean Busbey is correct. Even though nologin is the shell, you can still launch
processes as that user (and Accumulo runs as that user). I'm not sure if
Accumulo dumps stuff in the home directory, but a lot of other projects in
Bigtop do, so that's the standard practice. /var/run/... used to be used, but
on new Linux distributions that is a temporary FS and may be wiped on reboot.
Thanks for the review!
> Add Apache Accumulo
> -------------------
>
> Key: BIGTOP-1175
> URL: https://issues.apache.org/jira/browse/BIGTOP-1175
> Project: Bigtop
> Issue Type: Bug
> Reporter: Sean Mackrory
> Assignee: Sean Mackrory
> Attachments: 0001-BIGTOP-1175.-Add-Apache-Accumulo.patch
>
>
> I'd like to add Apache Accumulo to the stack (https://accumulo.apache.org)
--
This message was sent by Atlassian JIRA
(v6.2#6252)
