You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@helix.apache.org by GitBox <gi...@apache.org> on 2019/10/10 21:12:33 UTC
[GitHub] [helix] narendly opened a new issue #509: Allow multiple readers
for BucketDataAccessor
narendly opened a new issue #509: Allow multiple readers for BucketDataAccessor
URL: https://github.com/apache/helix/issues/509
Low Parallelism
Currently, ZkBasedBucketDataAccessor uses a locking mechanism implemented using ZooKeeper's ephemeral nodes. This is similar to a mutex, and allows for one entity to either read/write. Since there could only be one ongoing operation, the users of this data accessor such as WAGED rebalancer suffers from low parallelism. More specifically, it is not possible to have verifiers read from the metadata written to ZK while some RW is happening for testing/verification purposes.
Latency
Overall latency of the rebalance pipeline (BestPossibleCalcStage) could be reduced if we could parallelize read and write operations. The current implementation, due to locking, has to linearize all read and write operations.
Added Complexity from Locking
Inevitably, there will be unexpected failures. When we use locks in a distributed manner, we need to account for all failure scenarios related to connection issues. This makes coding a little more difficult because of many try-catch statements.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@helix.apache.org
For additional commands, e-mail: reviews-help@helix.apache.org