qubole / rubix

Cache File System optimized for columnar formats and object stores

qubole/rubix Issues

Support file size based usage limits
Updated 5 months ago
How can i join rubix slack
Updated 2 years ago1
Is there any document about the principle and architecture of rubix?
Updated 3 years ago1
Invalid cached data led to exceptions on read
Updated 4 years ago4
Avoid caching presto worker nodes
Updated 4 years ago2
Reactive disk usage accounting can lead to breaching cache size restriction for brief periods
Closed 4 years ago6
Publish jmx metrics for connection pool
Updated 4 years ago2
Prevent overwhelming of worker nodes by dynamically sizing thread pools
Updated 4 years ago2
Lower the default max wait timeout
Updated 4 years ago
Do not log `NoSuchFileException` as warning in `FileMetadata.addFilesForDeletion`
Updated 4 years ago
ClusterManager not updating cluster membership since 0.3.17
Closed 4 years ago
Lower log level of cache eviction to debug
Closed 4 years ago4
Lower DEFAULT_KEY_POOL_MAX_SIZE to smaller value, ideally under 10
Updated 4 years ago
Tests should run from the root directory
Closed 4 years ago1
Handle Interrupts better
Updated 4 years ago1
Use Stack in Connection-pool in-place of BlockingQueue
Updated 4 years ago
Space accounting issues in case of `rubix.cache.cleanup.files.during.start`=`false`
Updated 4 years ago
Avoid extra listStatus calls at worker level
Updated 4 years ago
Add option to disable non-local reads
Updated 4 years ago
Use `java.nio.file.Files` instead of `java.io.File`
Closed 4 years ago1
Available cache space is computed before cache directories are cleaned up during startup
Closed 4 years ago4
Fix jmx metrics documentation
Closed 4 years ago
Extend ClusterManager to return current node or current node index
Closed 4 years ago
Track moving averages in stats via airlift's CounterStat
Updated 4 years ago
Master Bookkeeper is getting overloaded with connection refused errors
Closed 4 years ago3
Use filesystem caching in servers
Closed 4 years ago1
Invalidations can cause corrupted reads
Closed 4 years ago3
Use factory pattern for DataTransferClientFactory to be passed from CachedFileSystem
Updated 4 years ago
Gracefully shutdown scanvenged connections at server side by sending a template close message.
Updated 4 years ago
Over-estimation of cache size for small files
Closed 4 years ago
Add timeouts to tests
Updated 4 years ago
DirectReadRequestChain and fallback-read might end up reading from same inputStream
Updated 4 years ago2
Bypass getFileInfo network call when staleness check is enabled
Closed 4 years ago1
Class not found: com.facebook.presto.hive.PrestoS3FileSystem
Closed 4 years ago8
Take cluster-type as config instead of setting it from client calls
Updated 4 years ago
S3 connection leak with PrestoS3FileSystem
Updated 4 years ago1
FileDownloadRequestChain can get ReadRequest of length > 2GB which causes integer overflow
Closed 4 years ago1
BufferUnderflowException in CachedReadRequestChain
Closed 4 years ago4
Add no-cache-on-master operation mode
Closed 4 years ago
Change block numbers to be of 'int' type instead of long
Updated 4 years ago
Bookeeper reports java.util.MissingFormatArgumentException instead root cause on read error
Closed 4 years ago1
Use single version of guava in all submodules
Closed 4 years ago
Travis UT should run with each PR
Closed 4 years ago
Refactor code to make BookkeeperFactory singleton
Updated 4 years ago
Refactor code to remove default constructor of RetryingPooledBookkeeperClient
Updated 4 years ago
Add owner tracing logic on Poolables
Updated 4 years ago
Move object pool to separate sub-module
Updated 4 years ago
Leak in ObjectPool when host goes away
Updated 4 years ago
r4
Closed 4 years ago1
Spark JVM not shutting down due to non-daemons threads
Closed 4 years ago1