[
https://issues.apache.org/jira/browse/HDFS-2802?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13486173#comment-13486173
]
Aaron T. Myers commented on HDFS-2802:
--------------------------------------
bq. Many of the things you are repeating again, about space and creation
efficiency etc. I have addressed them in many comments. Please do not comment
as if the current work is not considering it.
I definitely didn't mean to beat a dead horse on this subject. My apologies
that the comment came across that way. I was trying to enumerate the
differences between the documents, and propose a specific solution to this
problem. I certainly didn't intend to imply that the current work is not
considering this problem, hence why I referenced HDFS-4125 in my comment.
bq. Our solutions will be based on well known solution in the literature for
this problem. Please see on HDFS-4125.
I have been following HDFS-4125, but the comment in HDFS-4125 regarding the
solution you're referring to wasn't posted until after my most recent comment.
bq. Some of the differences around administrative can be made configurable, if
need be. These are surround aspect of the feature.
Sounds good.
bq. Given repeated feedback about optimizations we are working on updating the
implementation and design. Earlier the goal was to get the inefficient but
correct first implementation with tests and have design review, followed by
optimization. Given the feedback so far, we will post updated design next week.
Possibly hold a design review in around two weeks.
Cool. I look forward to reading the updated design doc.
bq. BTW if this design/implementation needs changes later we can have a
discussion. At this point in time, I would like to focus on getting parts of
the implementation we have had in our prototype out in this branch. I am fine
if we need to throw away parts of it.
My preference would be to collaborate on and agree to a design upfront so that
we can come up with something that will be amenable to everyone before we spend
a lot of time implementing something and potentially discarding it.
By the way, I forgot to mention that the design document I posted contains some
discussion and specific design proposals regarding the consistency of NN/DN
data in snapshots. Please take a look at that as well.
> Support for RW/RO snapshots in HDFS
> -----------------------------------
>
> Key: HDFS-2802
> URL: https://issues.apache.org/jira/browse/HDFS-2802
> Project: Hadoop HDFS
> Issue Type: New Feature
> Components: data-node, name-node
> Reporter: Hari Mankude
> Assignee: Hari Mankude
> Attachments: HDFSSnapshotsDesign.pdf, snap.patch,
> snapshot-one-pager.pdf, Snapshots20121018.pdf
>
>
> Snapshots are point in time images of parts of the filesystem or the entire
> filesystem. Snapshots can be a read-only or a read-write point in time copy
> of the filesystem. There are several use cases for snapshots in HDFS. I will
> post a detailed write-up soon with with more information.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
