Hello!
I have spent some time tinkering with your case, and I think that an issue
exists.
If you have a partitioned cache with non-local continuous query, it works.
If you have a replicated cache with non-local continuous query, it works.
If you have a replicated cache with local continuous query and synchronous
rebalancing, it also works.
But if you have replicated cache with local continuous query and
non-synchronous rebalancing (as you do), you will see the gap in data that
you are observing.
I still the best course of action in your case is to set,
RebalanceMode = CacheRebalanceMode.Sync
for "Test" cache.
However, I'm not 100% confident so I recommended running your test in a
loop for a couple minutes.
I will ponder if this is actually a bug, and maybe fill an issue against
Apache Ignite.
Regards,
--
Ilya Kasnacheev
пн, 1 окт. 2018 г. в 14:36, Alew <[email protected]>:
> Hi!
>
> Tried, but the issue still exists.
>
> https://monosnap.com/file/UUbYX4RUyXPZyPxKx97hTwSUSvNtye
>
> On 01/10/2018 13:06, Ilya Kasnacheev wrote:
>
> Hello!
>
> Setting BufferSize to 10000 seems to fix your reproducer's problem:
>
> new ContinuousQuery<string, byte[]>(new CacheListener(skippedItems,
> doubleDelete), true) { BufferSize = 10000}
>
> but the recommendation is to avoid doing anything synchronous from
> Continuous Query's body. Better offload any non-trivial processing to other
> threads operating asynchronously.
>
> Regards,
> --
> Ilya Kasnacheev
>
>
> сб, 29 сент. 2018 г. в 5:18, Alew <[email protected]>:
>
>> Hi, attached a reproducer.
>> Turn off logs to fix the issue. Slow log is not the only reason. More
>> nodes in a cluster lead to the same behaviour.
>> Who is responsible for the behavior? Is it .net, java, bad docs or me?
>>
>> On 24/09/2018 20:03, Alew wrote:
>> > Hi!
>> >
>> > I need a way to consistently get all entries in a replicated cache and
>> > then all updates for them while application is working.
>> >
>> > I use ContinuousQuery for it.
>> >
>> > var cursor = cache.QueryContinuous(new ContinuousQuery<string,
>> > byte[]>(new CacheListener(), true),
>> > new ScanQuery<string,
>> byte[]>()).GetInitialQueryCursor();
>> >
>> > But I have some issues with it.
>> >
>> > Sometimes cursor returns only part of entries in a cache and cache
>> > listener does not return them either.
>> >
>> > Sometimes cursor and cache listener return the same entry both.
>> >
>> > Issue somehow related to amount of work the nodes have to do and
>> > amount of time between start of the publisher node and subscriber node.
>> >
>> > There are more problems if nodes start at the same time.
>> >
>> > Is there a reliable way to do it without controling order of node
>> > start and pauses between them?
>> >
>> >
>>
>>
>
