-
Notifications
You must be signed in to change notification settings - Fork 3.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
HBASE-26907 Update Hadoop3 versions for JEP 223 compliance #4305
Conversation
See investigation and discussion on the JIRA. Maybe we want to drop 2.3 from the matrix in the book? Our personality makes no distinction in Hadoop3 versions for JDK8 vs. JDK11+. The latter requires 3.2+. We probably should fix that. |
This comment was marked as outdated.
This comment was marked as outdated.
2 similar comments
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
Well that's interesting. From hadoop-3.3.2:
|
we have our hbase-thirdparty no-op stand-in for htrace. We can add a profile for including that when building against some Hadoops? |
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Just a simple nit.
And on the compatibile matrix, IIRC we have a consensus that we will only have ref guide on master branch so it should contain all the release lines instead of only non-EOL release lines. So let's still keep the 2.3.x release line. Let me find the thread on mailing list and try to move forward on this topic.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM (agree with Duo's suggestion to bump to the latest 3.3.x)
@Apache9 okay ; I'll take your word for it, I can restore the 2.3 entries. Obviously we don't have all released mentioned, so at what point do we drop them off? |
I opened HBASE-26954 for the Hadoop 3.3.2 compilation issue. |
8a71d7c
to
5f8345d
Compare
I've bumped our Hadoop 3.2 dependency to 3.2.3, per @Apache9 's suggestion. I've retained the HBase 2.3 rows in the compatibility matrix. I've stepped back our promises about Hadoop 3.3.2, given HBASE-26954. |
Let me see which branches are impacted by HBASE-26954. |
HBASE-26954 seems to only affect |
This comment was marked as outdated.
This comment was marked as outdated.
2 similar comments
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This has not been well discussed yet, I mean when to drop the EOL versions. For me, I prefer we split the table to two pieces, for EOL versions, after some time(half a year or a whole year?), we move it to the 'archived' pieces, which means we will not update them any more. |
This comment was marked as outdated.
This comment was marked as outdated.
Any updates here? |
Progress is reported over on the linked ticket where I describe the test failure. There's a ThreadGroup object that goes out of scope during DataNode operation on these newer Hadoop versions. I think that the explicitly managed ThreadGoup serves no purpose, but I'm still looking for a commit that changed the behavior. I believe is indicates a real and serious bug in the DataNode, which will cause it to become unresponsive after an idle period. |
I think @saintstack fixed the issue i was seeing in HDFS-16586 / apache/hadoop#4338 . Let me see how widely that patch has propagated. |
Ah, I've replied on the jira, in HBASE-27148 we have already moved up the hadoop 3.2.3. I also noticed the problem described in HDFS-16586 but I'm afraid we can not only support the hadoop versions which have this patch since it is only a test issue. I added a piece of workaround code in HBTU so we can still use the versions which do not have HDFS-16586 in.
So I think we could close this PR now? |
0b9fb7d
to
a215a0a
Compare
Rebased and squashed current patched onto latest master. |
@Apache9 Ah okay, I missed HBASE-27148. |
It looks like we still have hadoop-3.1.2 in the quick-check list for branch-2.4, I guess we can bump that version. Also, it appears that I have added a column regarding branch-2.5 in this patch. Let me push an updated based the knowledge of HBASE-27148. |
Sounds good. And please see #4692 , where I plan to change the hadoop check versions for our active branches, which also implies what hadoop versions we still wantto support. I think for 2.4.x, we should not remove the support of a hadoop version, but for 2.5+, we can drop the EOL versions and also the versions which have critical CVEs. |
- drop 3.1.x - ensure 3.2 >= 3.2.3 - add 3.3.3, 3.3.4
(!) A patch to the testing environment has been detected. |
2 similar comments
(!) A patch to the testing environment has been detected. |
(!) A patch to the testing environment has been detected. |
🎊 +1 overall
This message was automatically generated. |
🎊 +1 overall
This message was automatically generated. |
Hmm okay. @Apache9 how about I close HBASE-26907 this PR and let you resolve them on HBASE-27294. Do you might grabbing the update to the book once that ticket is resolved? Thanks. |
For the book update let's open a separated issue? As part of the branch-2.5 release process. |
🎊 +1 overall
This message was automatically generated. |
Ah, maybe I misunderstood your comment. You also mean we open a new issue to land the book update? That's fine by me. I could help implementing the PR or reviewing the PR, no problem. Thanks. |
(!) A patch to the testing environment has been detected. |
2 similar comments
(!) A patch to the testing environment has been detected. |
(!) A patch to the testing environment has been detected. |
🎊 +1 overall
This message was automatically generated. |
🎊 +1 overall
This message was automatically generated. |
🎊 +1 overall
This message was automatically generated. |
Superceded HBASE-27148, HBASE-27294. |
No description provided.