Hibernate Search integration modules for WildFly have been broken
by Sanne Grinovero
Hi all,
I was going to give ISPN-8779 a shot but I'm finding a mess.
the root pom contains these twice (and inconsistent!):
<version.hibernate.search>5.8.1.Final</version.hibernate.search>
[...]
<version.hibernate.search>5.8.0.Final</version.hibernate.search>
the BOM cointains a copy of `version.hibernate.search` as well.
I don't mind deleting duplicate properties, but we used to have
clearly separate properties for different purposes, and this
separation is essential.
I've mentioned this multiple times when reviewing PRs which would get
my attention, but I didn't see these changes - certainly didn't expect
you all to forget the special purpose of these modules.
It's quite messy now and I'm honestly lost myself at how I could revert it.
In particular this module is broken now as it's targeting the wrong slot:
- https://github.com/infinispan/infinispan/blob/master/wildfly-modules/src/...
Clearly it's not consistent with the comment I've put on the module descriptor.
I don't see that module being included in the released modules either,
and clearly the integration tests didn't catch it because they have
been patched to use the wrong modules too :(
Other essential integration tests which I had put in place to make
sure they'd get your attention in case someone had such an idea.. have
been deleted.
Opening ISPN-8780, I would consider this a release blocker.
Thanks,
Sanne
See also:
- https://github.com/infinispan/infinispan/blob/master/integrationtests/as-...
6 years, 2 months
Best practices for Netty version clashes
by Galder Zamarreño
Hi,
I was playing around with GRPC for a talk next month and made a mistake
that threw me a little bit and wanted to share it here to see if we can
do something about it.
My demo uses GRPC and Infinispan embedded cache (9.2.0.CR1), so I added
my GRPC dependencies and Infinispan bom dependency [1].
This combo resulted in breaking my GRPC demos.
The bom imports Netty 4.1.9.Final while GRPC requires 4.1.17.Final. The
dependency tree showed GRPC using 4.1.9.Final which lead to the
failure. This failure does not seem present in 4.1.17.Final.
Should we have an embedded bom where no client libraries are depended
upon? This would work for my particular use case...
However, someone might develop a GRPC server (which I *think* it still
requires netty) and they could then use Infinispan remote client to
bridge over to Infinispan sever. For example: this could be way to move
clients over a new client while other clients use an older protocol.
How should a user solve this clash? I can only see exclusions and
depending on latest Netty version as solution. Any other solutions
though?
Cheers,
[1] https://gist.github.com/galderz/300cc2708eab76b9861985c216b90136
6 years, 2 months
ISPN-8798 ByteString places too strict a constraint on cache name length
by Paul Ferraro
Can one of the devs please review this patch?
https://github.com/infinispan/infinispan/pull/5750
The limit of cache names sizes to 127 bytes is too limiting for
hibernate/JPA 2nd level cache deployments, which generate cache names
using fully qualified class names of entity classes, which are user
generated thus can easily exceed 128 bytes (but are far less likely to
exceed 255). This is exacerbated by the JPA integration, which
additionally appends the deployment name. We have a long term
solution for this, but in the meantime, the above patch is sufficient
to pass the TCK.
We'll also need a 9.1.6.Final release ASAP, lest we revert back to
Infinispan 8.2.x for WF12, the feature freeze for which is tomorrow
(they are considering this upgrade a feature, given the scope of its
impact).
Thanks,
Paul
6 years, 2 months
9.2.0 endgame plan
by Tristan Tarrant
I had originally planned for a release for Wed 14th, but there are a
number of things I'd like to see landing before Final and looking at the
list I recommend doing a CR3.
In particular:
- Radim's Hot Rod changes
- Performance regressions as reported by Will
- Ensure that Sanne is happy with the WF modules
- Documentation and quickstarts/simple tutorials for new features
- Quickstarts/simple tutorials work flawlessly
Tristan
P.S.
I'll be on PTO Mon/Tue 12/13 February.
--
Tristan Tarrant
Infinispan Lead and Data Grid Architect
JBoss, a division of Red Hat
6 years, 2 months
PersistentUUIDManagerImpl NPEs being logged when running the testsuite
by Sanne Grinovero
Hi all,
I'm building master [1] and see such NPEs dumped on my terminal quite
often; I guess you all noticed already? I couldn't find a JIRA..
16:24:03,083 FATAL
(transport-thread-StateTransferLinkFailuresTest[null,
tx=false]-NodeN-p63985-t2) [PersistentUUIDManagerImpl] Cannot find
mapping for address StateTransferLinkFailuresTest[null,
tx=false]-NodeN-32100 java.lang.NullPointerException
at org.infinispan.topology.PersistentUUIDManagerImpl.mapAddresses(PersistentUUIDManagerImpl.java:70)
at org.infinispan.partitionhandling.impl.PreferAvailabilityStrategy.onPartitionMerge(PreferAvailabilityStrategy.java:214)
at org.infinispan.topology.ClusterCacheStatus.doMergePartitions(ClusterCacheStatus.java:597)
at org.infinispan.topology.ClusterTopologyManagerImpl.lambda$recoverClusterStatus$6(ClusterTopologyManagerImpl.java:519)
at org.infinispan.executors.LimitedExecutor.runTasks(LimitedExecutor.java:144)
at org.infinispan.executors.LimitedExecutor.access$100(LimitedExecutor.java:33)
at org.infinispan.executors.LimitedExecutor$Runner.run(LimitedExecutor.java:174)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
16:24:03,115 FATAL
(transport-thread-StateTransferLinkFailuresTest[null,
tx=false]-NodeQ-p64193-t5) [PersistentUUIDManagerImpl] Cannot find
mapping for address StateTransferLinkFailuresTest[null,
tx=false]-NodeQ-10499 java.lang.NullPointerException
at org.infinispan.topology.PersistentUUIDManagerImpl.mapAddresses(PersistentUUIDManagerImpl.java:70)
at org.infinispan.partitionhandling.impl.PreferAvailabilityStrategy.onPartitionMerge(PreferAvailabilityStrategy.java:214)
at org.infinispan.topology.ClusterCacheStatus.doMergePartitions(ClusterCacheStatus.java:597)
at org.infinispan.topology.ClusterTopologyManagerImpl.lambda$recoverClusterStatus$6(ClusterTopologyManagerImpl.java:519)
at org.infinispan.executors.LimitedExecutor.runTasks(LimitedExecutor.java:144)
at org.infinispan.executors.LimitedExecutor.access$100(LimitedExecutor.java:33)
at org.infinispan.executors.LimitedExecutor$Runner.run(LimitedExecutor.java:174)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
1 - cc2744e9f509d917f1ed0ff1a18b28b72595af83
Thanks,
Sanne
6 years, 3 months
