[
https://issues.apache.org/jira/browse/HDFS-5535?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13912160#comment-13912160
]
Andrew Wang commented on HDFS-5535:
-----------------------------------
Hi Suresh, thanks for commenting back:
bq. <longer upgrades>
This is something Ming Ma asked about above regarding TOR switch upgrades, and
Kihwal said it might be considered with lite-decom. Basically, what to do when
the restart will take longer than the client timeout. Doing a full decom in
this case sucks since we know the DN will be back soon, but just not soon
enough to wait it out. I was wondering if this was going to be handled, now
that you're getting close with the rest.
bq. <compatibility and downgrade>
If we get feature flags finished (HDFS-5223), we can keep edit/fsimage
compatibility between minor releases until the operator enables a new feature.
Even after it's enabled, like you said, we'll sometimes be able to just drop
the new information on downgrade, e.g. ignore ACLs on an inode.
I agree that sometimes we'll have to break compat, but rolling upgrade +
feature flags + PB means our upgrade/downgrade story is better for operators
who are primarily concerned with bug fixes rather than new features.
> Umbrella jira for improved HDFS rolling upgrades
> ------------------------------------------------
>
> Key: HDFS-5535
> URL: https://issues.apache.org/jira/browse/HDFS-5535
> Project: Hadoop HDFS
> Issue Type: New Feature
> Components: datanode, ha, hdfs-client, namenode
> Affects Versions: 3.0.0, 2.2.0
> Reporter: Nathan Roberts
> Attachments: HDFSRollingUpgradesHighLevelDesign.pdf,
> h5535_20140219.patch, h5535_20140220-1554.patch, h5535_20140220b.patch,
> h5535_20140221-2031.patch, h5535_20140224-1931.patch,
> h5535_20140225-1225.patch
>
>
> In order to roll a new HDFS release through a large cluster quickly and
> safely, a few enhancements are needed in HDFS. An initial High level design
> document will be attached to this jira, and sub-jiras will itemize the
> individual tasks.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)