Repeating my +1.
1. Deployed the branch to k8s based cluster.
2. Executed Spark job to generate parquet file.
3. Deleted / re-created one datanode
4. Checked the if the files are recovered after a few minutes
5. Executed Spark job to read (and copy) the generated parquet files
Worked well.
Marton
ps: And thanks again to drive this Stephen. Good to see that this branch
will be merged finally...
On 11/12/20 4:46 PM, Stephen O'Donnell wrote:
Following on from the discussion thread, I would like to call a vote on
whether we should merge HDDS-1880 into master. This feature will enable
decommission and maintenance mode for datanodes. The feature is working,
but more work is needed to improve the usability etc. I feel the feature is
stable and self contained enough to continue that work on the master branch
now
The overall design is under HDDS-1880 and there is a summary of the
internals also attached to that Jira.
A PR for the merge is here:
https://github.com/apache/ozone/pull/1580
I will try to keep the PR up-to-date if any conflicts arise.
From the discussion thread, I can carry over 1 +1 from Marton Elek to get
us started.
As it may take some time to review the large PR I will let the vote run for
14 days, closing at midnight GMT on 26th November.
Please ask any questions here, on the PR or on ASF Slack.
Thanks,
Stephen.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@ozone.apache.org
For additional commands, e-mail: dev-h...@ozone.apache.org
