With 4 binding +1, 1 non-binding +1, and no binding -1, the vote passes!
Thank you everyone for voting.
I want to give a special thanks to Dieter for raising his non-binding -1.
The vote is acknowledged, and we will try to prioritize reviewing & merging
the issues he raised so that we can do a
Here's my own binding +1:
* Signature: ok
* Checksum : ok
* Rat check (11.0.18): ok
- mvn clean apache-rat:check
* Built from source (11.0.18): ok
- mvn clean install -DskipTests
* Unit tests pass (11.0.18): ok
- mvn
+1 (binding)
* Signature: ok
* Checksum : ok
* Rat check (1.8.0_412): ok
- mvn clean apache-rat:check
* Built from source (1.8.0_412): ok
- mvn clean install -DskipTests
* Built from source (11.0.23): ok
- mvn clean install -D hadoop.profile=3.0 -DskipTests
* Built from source (17.0.11): ok
Related, Dieter, I'm going to try to rally some review bandwidth on the
jiras you have linked. I really appreciate your team's efforts in reporting
and providing fixes for these issues. Some of them slipped past me, but I'm
now watching all of them and hope to get to them soon.
On Tue, May 7,
Yes, at my employer we are running daily backups using the new system,
including incremental backups, on ~130 clusters including 2 production
clusters. We've had troubles with incremental backups on a few clusters,
but most have been fine. We've also got end-to-end recurring acceptance
tests which
The new backups system is one of the "flagship" features for the 2.6
release line so it's a shame that these issues with incremental manifest
tracking remain. If I'm not mistaken, though, a full backup still works as
expected, so these bugs do not prevent taking a backup entirely. I agree
that
This sounds very reasonable to me, especially with the promise of a quick
follow on release of 2.6.1.
—
I apologize for not voting on the 2.6 candidates. I will set up a VM
somewhere where I can drive release candidate tests by phone for the next
time but did not think of this in advance.
On
I'm +1 on release 2.6.0 first.
For a new minor release line, we could do more aggressive patch
releasing in the beginning if there are lots of commits get in. For
example, if we could get 20+ commits soon, we could do 2.6.1 release
after two weeks.
Bryan Beaudreault 于2024年5月7日周二 18:54写道:
>
>
Thanks Dieter. I’m aware of these issues, but I don’t think they are
blockers. The idea with releasing backups was that it would be experimental
in 2.6, as a way to get the feature into more of the community’s hands and
increase development. So far that’s working, as evidenced by your team’s
great
-1 non binding
I've done some testing of the backup-restore feature the past days, and there's
still some issues that I think really should be solved for a first release
version.
PRs are available for all of these:
* HBASE-28539: backup merging does not work when using cloud storage as
+1 non-binding
- Changes, Releasenotes, Signature: Ok
- Building from source JDK 17 : Ok
- Rat check : Ok
- Unit test: Ok
- Performed migration from 2.5.2 to 2.6.0: Ok.
- Ran basic sanity of CRUD, snapshots, PE, bulkloads: Ok
Thanks,
Rajeshbabu.
On Thu, May 2, 2024 at 6:29 PM Peter Somogyi
+1 binding
* changes, release notes: ok
* Signature: ok
* Checksum : ok
* Rat check (17.0.11): ok
- mvn clean apache-rat:check -D hadoop.profile=3.0
* Built from source (17.0.11): ok
- mvn clean install -D hadoop.profile=3.0 -DskipTests
+1 binding, will go out in the next several days so let me vote now.
Checked sigs and sums: Matched
Rat check: Passed
LICENSE and NOTICE: In place
Compatibility report:
Most of incompatibility are from thirdparty jars like zookeeper, the
only one is RegionInfo.UNDEFINED, which is IA.Private so
Please vote on this Apache hbase release candidate,
hbase-2.6.0RC4
The VOTE will remain open for at least 72 hours.
[ ] +1 Release this package as Apache hbase 2.6.0
[ ] -1 Do not release this package because ...
The tag to be voted on is 2.6.0RC4:
14 matches
Mail list logo