[
https://issues.apache.org/jira/browse/HDDS-839?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16687784#comment-16687784
]
Elek, Marton commented on HDDS-839:
-----------------------------------
Patch is uploaded. To test it:
1. Apply the patch to the docker-hadoop-runner branch.
2. cd scripts
3. Test the waiting
In one terminal:
{code}
WAITFOR=localhost:1234 ./starter.sh ls
{code}
In an other terminal:
{code}
nc -l -p 1234
{code}
Script should continue only after the nc server is started
4. the normal case (without waiting)
{code]
./starter.sh ls
{code}
You can see the output immediately
5. Timeout
{code}
WAITFOR_TIMEOUT=10 WAITFOR=localhost:1234 ./starter.sh ls
{code}
You can see an error message (and 1 exit code) after ~10 seconds (unless you
start a server on port 1234)
> Wait for other services in the started script of hadoop-runner base docker
> image
> --------------------------------------------------------------------------------
>
> Key: HDDS-839
> URL: https://issues.apache.org/jira/browse/HDDS-839
> Project: Hadoop Distributed Data Store
> Issue Type: Sub-task
> Reporter: Elek, Marton
> Assignee: Elek, Marton
> Priority: Major
>
> As described in the parent issue, we need a simple method to handle service
> dependencies in kubernetes clusters (usually as a workaround when some
> clients can't re-try with renewed dns information).
> But it also could be useful to minimize the wait time in the docker-compose
> clusters.
> The easiest implementation is modifying the started script of the
> apache/hadoop-runner base image and add a bash loop which checks the
> availability of the TCP port (with netcat).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
