-
Notifications
You must be signed in to change notification settings - Fork 698
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
SOLR-17012: Update Apache Hadoop to 3.3.6 and Apache Curator to 5.5.0 #1743
Conversation
c02e998
to
7846230
Compare
I've given up hope for 3.4.0 at this point. |
7846230
to
cb7235b
Compare
cb7235b
to
465c426
Compare
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.
need to check some of the test results...
https://github.com/apache/hadoop/blob/rel/release-3.3.6/hadoop-project/pom.xml#L103 hadoop upgraded to curator 5.x - https://issues.apache.org/jira/browse/HADOOP-18515 |
#1427 is handled here |
The smaller subset of errors after upgrading to curator 5.2.0
|
Edited/Blocked NotificationRenovate will not automatically rebase this PR, because it does not recognize the last commit author and assumes somebody else may have edited the PR. You can manually request rebase by checking the rebase/retry box above. ⚠ Warning: custom changes will be lost. |
…ace conditions to create znodes...
…re are race conditions to create znodes..." This reverts commit 40a8228.
Some of the Hadoop test failures were just normal thread leaks that were handled by de729bb There were another subset of failures that were more interesting. I found a solution to the Hadoop test failures: 40a8228 The failure was that Solr through Hadoop's There is probably a fix in
|
solr/test-framework/src/java/org/apache/solr/cloud/MiniSolrCloudCluster.java
Outdated
Show resolved
Hide resolved
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.
The changes look good. We may not know for certain if there are perf implications on a single threaded executor here but given the support burden, I'd rather err on the side of our convenience.
The change in Curator version from 4x to 5x means losing compatibility with ZooKeeper 3.4 -- which is fine. But I think it should be documented in the upgrade notes. Obviously it'd only apply to the (small?) subset of users actually using Hadoop-auth.
https://curator.apache.org/zk-compatibility-34.html. (it refers to Curator 4.2 but I believe equally applies to later versions in 4x).
Argggg same exception with some nightly tests. It just dumb that there is this race condition. I'm poking around still to see if there is a way forward. |
Sigh so I found apache/hadoop#4885 which tried to fix the problem but the race condition is still there :/ I think the fix should be to catch the exception if the znode exists. |
So I reverted the single threaded portion since there are other code paths that spin up Solr in parallel. I implemented ae141d8 instead which creates the znode the same way as Hadoop upfront. Before Hadoop has a chance to check. This will properly ignore if the znode already exists. |
I'm running tests and nightly tests hopefully this is a reasonable step forward. |
nightly and regular tests all passed for me now. |
re: upgrade notes - https://solr.apache.org/guide/solr/latest/upgrade-notes/major-changes-in-solr-9.html#solr-8-2 has a note about zookeeper 3.5 but happy to put a note in as well. |
upgraded to latest curator 5.5.0 since it had some decent bug fixes (no major features). only thing left is CHANGES and upgrade notes about curator 5.x and zookeeper 3.4 |
I created https://issues.apache.org/jira/browse/SOLR-17012 since other work depends on curator being 5.x so wanted to call it out more. |
…#1743) Co-authored-by: Kevin Risden <[email protected]>
The existing notes don't address my point at all. And as we embrace Curator (in a separate issue), the lack of ZK 3.4 compatibility will become more pronounced (not limited to Hadoop Auth). |
agreed I added new upgrade notes was just linking to the point that we had mentioned ZK 3.5 recommended a while ago so hopefully 3.4 isn't as big of an issue. |
Is it worth having a ZK matrix in ref guide similar to the sections on Java in https://solr.apache.org/guide/solr/latest/deployment-guide/system-requirements.html#released-solr-and-java-versions ? I've seen in various forums ask "Is my ZK too old to work with Solr X" |
FWIW I put together https://issues.apache.org/jira/browse/HADOOP-18922 and apache/hadoop#6150 to address the race condition upstream. |
…#1743) Co-authored-by: Kevin Risden <[email protected]>
This PR contains the following updates:
3.3.5
->3.3.6
3.3.5
->3.3.6
3.3.5
->3.3.6
3.3.5
->3.3.6
3.3.5
->3.3.6
3.3.5
->3.3.6
3.3.5
->3.3.6
3.3.5
->3.3.6
Configuration
📅 Schedule: Branch creation - "before 3am on the first day of the month" (UTC), Automerge - At any time (no schedule defined).
🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.
♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.
🔕 Ignore: Close this PR and you won't be reminded about these updates again.
This PR has been generated by Renovate Bot