> On Feb. 2, 2017, 10:45 p.m., Jason Huynh wrote:
> > geode-lucene/src/main/java/org/apache/geode/cache/lucene/internal/LuceneBucketListener.java,
> > line 58
> > <https://reviews.apache.org/r/56244/diff/1/?file=1622552#file1622552line58>
> >
> > Should we create a ticket to change this to a specific list of
> > exceptions we may be expecting?
should this be similar to the behaviour of afterPrimary ?
try {
lucenePartitionRepositoryManager.computeRepository(bucketId);
} catch (BucketNotFoundException e) {
logger.warn(
"Index repository could not be created when index chunk region
bucket became primary. "
+ "Deferring index repository to be created lazily during
lucene query execution."
+ e);
}
- nabarun
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/56244/#review164038
-----------------------------------------------------------
On Feb. 2, 2017, 10:32 p.m., nabarun nag wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/56244/
> -----------------------------------------------------------
>
> (Updated Feb. 2, 2017, 10:32 p.m.)
>
>
> Review request for geode, Barry Oglesby, Jason Huynh, Lynn Hughes-Godfrey,
> Dan Smith, and xiaojian zhou.
>
>
> Repository: geode
>
>
> Description
> -------
>
> Design key points:
> ==================
> 1. The afterPrimary and afterSecondary calls from the chunk bucket pass
> through the same critical section
> 2. If the bucket is still primary it will attempt to acquire a Dlock on the
> bucket and create the index repo.
> 3. The primary call tries for 5 seconds to acquire the Dlock and then checks
> if it is still primary. If it is, it will retry to acquire the lock. If it
> turned into a secondary it will exit the critical section. It is assumed that
> the after secondary call will do the cleanup.
> 4. The afterSecondary call will simply do the clean up - close the writer,
> release locks, undo contributions to the lucene stats.
> 5. Also, in a situation when index creation fails during after primary,
> lucene query execution will try to recompute the index once again.
>
>
> Diffs
> -----
>
>
> geode-lucene/src/main/java/org/apache/geode/cache/lucene/internal/AbstractPartitionedRepositoryManager.java
> 9e055f0
>
> geode-lucene/src/main/java/org/apache/geode/cache/lucene/internal/IndexRepositoryFactory.java
> 5be17e3
>
> geode-lucene/src/main/java/org/apache/geode/cache/lucene/internal/LuceneBucketListener.java
> da0c2c2
>
> geode-lucene/src/main/java/org/apache/geode/cache/lucene/internal/LuceneQueryImpl.java
> 0e8bb37
>
> geode-lucene/src/main/java/org/apache/geode/cache/lucene/internal/PartitionedRepositoryManager.java
> 2f87218
>
> geode-lucene/src/main/java/org/apache/geode/cache/lucene/internal/RawIndexRepositoryFactory.java
> 2f61913
>
> geode-lucene/src/main/java/org/apache/geode/cache/lucene/internal/RawLuceneRepositoryManager.java
> b503692
>
> geode-lucene/src/main/java/org/apache/geode/cache/lucene/internal/directory/RegionDirectory.java
> e43b60b
>
> geode-lucene/src/main/java/org/apache/geode/cache/lucene/internal/distributed/LuceneFunction.java
> 5271a2f
>
> geode-lucene/src/test/java/org/apache/geode/cache/lucene/internal/PartitionedRepositoryManagerJUnitTest.java
> 1c47e89
>
> geode-lucene/src/test/java/org/apache/geode/cache/lucene/internal/RawLuceneRepositoryManagerJUnitTest.java
> a9fb52b
>
> geode-lucene/src/test/java/org/apache/geode/cache/lucene/internal/directory/RegionDirectoryJUnitTest.java
> 4204204
>
> geode-lucene/src/test/java/org/apache/geode/cache/lucene/test/IndexRepositorySpy.java
> d3b1f2f
>
> Diff: https://reviews.apache.org/r/56244/diff/
>
>
> Testing
> -------
>
> * Lucene precheck successfull
> * Hydra failures are imminent - work ongoing on EOF + rebalance exceptions
>
>
> Thanks,
>
> nabarun nag
>
>
