You are viewing a plain text version of this content. The canonical link for it is here.
Posted to hdfs-dev@hadoop.apache.org by "Elek, Marton (JIRA)" <ji...@apache.org> on 2018/02/28 08:28:00 UTC

[jira] [Created] (HDFS-13203) Ozone: dozone: auto-bootstrap SCM/KSM in containerized environment

Elek, Marton created HDFS-13203:
-----------------------------------

             Summary: Ozone: dozone: auto-bootstrap SCM/KSM in containerized environment
                 Key: HDFS-13203
                 URL: https://issues.apache.org/jira/browse/HDFS-13203
             Project: Hadoop HDFS
          Issue Type: Sub-task
          Components: HDFS-7240
    Affects Versions: 3.1.0
            Reporter: Elek, Marton
            Assignee: Elek, Marton


Currently to start SCM and KSM in a new environment we need to:

 1. Initialize scm (/hdfs scm -init)
 2. Start scm
 3. hdfs ksm -createObjectStore (connects to the scm, scm should be running)
 4. Start ksm

The problematic part is step 3.) as in a containerized environment (docker-compose/kubernetes) the timing couldn't be guaranteed.

I propose to introduce a new command line flag '--auto-bootstrap' for both KSM and SCM (and maybe for Namenode in the future).

It will:
  1. Initilize KSM and SCM *if and only if* the data directories are missing
  2. In case of KSM retry the initilization in every 10 seconds (in case of SCM is not yet running) in case of connection failure



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-dev-unsubscribe@hadoop.apache.org
For additional commands, e-mail: hdfs-dev-help@hadoop.apache.org