ISPN-14192 Do not send empty batches with IRAC #10367
Conversation
commons/all/src/main/java/org/infinispan/commons/util/ConcurrentSmallIntSet.java
Outdated
Show resolved
Hide resolved
| @@ -178,7 +178,8 @@ public static IntSet mutableSet(int value1, int value2) { | |||
| * @return concurrent set | |||
| */ | |||
| public static IntSet concurrentSet(int maxExclusive) { | |||
| return new ConcurrentSmallIntSet(maxExclusive); | |||
| // if maxExclusive = 0; then we have an empty set | |||
| return maxExclusive < 1 ? immutableEmptySet() : new ConcurrentSmallIntSet(maxExclusive); | |||
There was a problem hiding this comment.
I would personally say only return immutable empty set when maxExclusive == 0. That way we will get the new exception you added above if it is negative.
| @@ -189,6 +190,10 @@ public static IntSet concurrentSet(int maxExclusive) { | |||
| * @return concurrent copy | |||
| */ | |||
| public static IntSet concurrentCopyFrom(IntSet intSet, int maxExclusive) { | |||
| // if maxExclusive = 0; then we have an empty set | |||
| if (maxExclusive < 1) { | |||
There was a problem hiding this comment.
I would personally say only return immutable empty set when maxExclusive == 0. That way we will get the new exception you added above if it is negative.
| } | ||
| onUnexpectedThrowable(throwable); | ||
| return Completable.complete(); |
There was a problem hiding this comment.
This slightly changes how exception handling operates. With the new changes if there is a Throwable, we now have to wait for the registered remote site backup commands to complete. Do we want that? It is probably a very small chance of this happening, but just wanted to make sure.
There was a problem hiding this comment.
Do you mean the invalidState list being cleaned up now and not before?
There is no reason to keep the invalid state around more time than necessary to be cleaned up in the "next round".
There was a problem hiding this comment.
Not quite, If you have more than one backup if the second errors, we would have to wait until the first completes as rspCollector would have a stage dependency.
There was a problem hiding this comment.
I see. I'm setting rspCollector = null; to avoid waiting although, I think, this code will never be executed.
There was a problem hiding this comment.
Oh, yes I missed that sorry. So we are good.
* Includes change to ConcurrentSmallIntSet preventing wrong constructor argument
https://issues.redhat.com/browse/ISPN-14192