There is a fix in PR#391.

I've run the test case umpteen times now without incident and done performance times of bulkloading, which is about as write intensive as it gets.

The worse slow down I observed (difference of a 3.6 and 3.7-fix) when loading 25M BSBM triples was 3%. Usually it's 1%. Larger loads will be dominated by I/O costs so what is a CPU-related fix will be overtaken (I hope!). It did in the testing I did.

So this email cancels RC1 and I'll start again.

     Andy

On 03/04/18 23:01, Andy Seaborne wrote:
JENA-1516 looks serious and fixable.

I'd like time to verify the fix that I have. As is often the case, finding the concurrency problem is a lot more time consuming than fixing an identified problem.

The WIP fix is available at
   https://github.com/afs/jena/tree/jena-1516_ObjectFileStorage
if you are really interested.

If that fix looks good, I propose cancelling this RC and doing a new one to include the fix.

If the fix does not look good, and it has taken quite a heavy sustained concurrent load to get anything to go wrong (carefully parallel 100 updates, 100 queries, 50 threads and it still does not happen every time), it seems better to continue RC1 and get a good fix done.

Adam, Chris - thank you for the time to assess RC1.  There should be very little change between RC1 and a possible RC2.

(not sure if SUSPENDED is a recognized state for a VOTE - think of it as the RM needs time to do some more checking)

     Andy

On 29/03/18 19:28, Andy Seaborne wrote:
Hi,

Here is a vote on a release of Jena 3.7.0.

This is the first proposed candidate for a 3.7.0 release.

There are process changes.

Deadline:

        2018-04-01 22:00 UTC

April 1st!

==== Process Changes

1/
MD5 files are being discouraged because MD5 is not secure.  Projects are now asked to not publish md5.

There are no md5 files in the proposed dist/jena area - files on Apache hardware.

There are sha1 and sha512 checksums.
* The sha512 is in Linux sha512sum checkable format.
* The sha1 is whatever maven generated and is the same as will go to maven central.

Having the sha1 ties the dist/jena artifacts to maven central (as does the .asc).

There are md5 and sha1 in the proposes maven repo staging area for sending to maven central. That part of maven is hardwired to md5/sha1 still.

There's a script to setup the sha512.

2/
To establish the proof chain for signed artifacts in /dist/project/, I have been asked to try out the new META files.

https://checker.apache.org/doc/README.html#ch-meta

There are two files

/dist/jena/META
/dist/jena/META.asc

META says who signs what, and is itself signed by the PMC chair.

==== Release changes

55 JIRA:
https://s.apache.org/jena-3.7.0-jira

== Significant Changes

** Java9: Building and running on a Java9 platform is supported

JENA-1461 - Allow ARQ custom functions to be written in JavaScript

JENA-1389 - Return `this` rather than `void` from Dataset (API change)
JENA-1495 - Return Model from PrefixMapping methods (API change)

JENA-1458, JENA-1483 - Transaction Promotion

JENA-1453 - Lucene indexes using a graph field are smaller

JENA-1490 - Working with Blank Nodes with Fuseki

== Upgrades to libraries (runtime dependencies):

No dependency changes.

==== Release Vote

Everyone, not just committers, is invited to test and vote.
Please download and test the proposed release.

Proposed dist/ area:
   https://dist.apache.org/repos/dist/dev/jena/

Keys:
   https://svn.apache.org/repos/asf/jena/dist/KEYS

Staging repository:
   https://repository.apache.org/content/repositories/orgapachejena-1022/

Git commit (browser URL):
https://git1-us-west.apache.org/repos/asf?p=jena.git;a=commit;h=d4e7063e

Git Commit Hash:
   d4e7063e7a6db8ce77699bd0388e1a1bd6816626

Git Commit Tag:
      jena-3.7.0-rc1

Please vote to approve this release:

        [ ] +1 Approve the release
        [ ]  0 Don't care
        [ ] -1 Don't release, because ...

This vote will be open until at least

        2018-04-01 22:00 UTC

If you expect to check the release but the time limit does not work
for you, please email within the schedule above with an expected time
and we can extend the vote period.

Thanks,

      Andy

Checking needed:

+ does everything work on Linux?
+ does everything work on MS Windows?
+ does everything work on OS X?
+ are the GPG signatures fine?
+ are the checksums correct?
+ is there a source archive?

+ can the source archive really be built?
          (NB This requires a "mvn install" first time)
+ is there a correct LICENSE and NOTICE file in each artifact
          (both source and binary artifacts)?
+ does the NOTICE file contain all necessary attributions?
+ have any licenses of dependencies changed due to upgrades?
           if so have LICENSE and NOTICE been upgraded appropriately?
+ does the tag/commit in the SCM contain reproducible sources?

Reply via email to