[HBASE-11425] Cell/DBB end-to-end on the read-path

• Type: Umbrella
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.0.0
• Component/s: regionserver, Scanners
• Hadoop Flags: Reviewed

Release Note

For E2E off heaped read path, first of all there should be an off heap backed BucketCache(BC). Configure 'hbase.bucketcache.ioengine' to offheap in hbase-site.xml. Also specify the total capacity of the BC using hbase.bucketcache.size config. Please remember to adjust value of 'HBASE_OFFHEAPSIZE' in hbase-env.sh as per this capacity. Here-by we specify the max possible off-heap memory allocation for the RS java process. So this should be bigger than the off-heap BC size. Please keep in mind that there is no default for hbase.bucketcache.ioengine which means the BC is turned OFF by default.

Next thing to tune is the ByteBuffer pool in the RPC server side. The buffers from this pool will be used to accumulate the cell bytes and create a result cell block to send back to the client side. 'hbase.ipc.server.reservoir.enabled' can be used to turn this pool ON or OFF. By default this pool is ON and available. HBase will create off heap ByteBuffers and pool them. Please make sure not to turn this OFF if you want E2E off heaping in read path. If this pool is turned off, the server will create temp buffers on heap to accumulate the cell bytes and make a result cell block. This can impact the GC on a highly read loaded server. The user can tune this pool with respect to how many buffers are in the pool and what should be the size of each ByteBuffer.
Use the config 'hbase.ipc.server.reservoir.initial.buffer.size' to tune each of the buffer sizes. Defaults is 64 KB.

When the read pattern is a random row read and each of the rows are smaller in size compared to this 64 KB, try reducing this. When the result size is larger than one ByteBuffer size, the server will try to grab more than one buffer and make a result cell block out of these. When the pool is running out of buffers, the server will end up creating temporary on-heap buffers.

The maximum number of ByteBuffers in the pool can be tuned using the config 'hbase.ipc.server.reservoir.initial.max'. Its value defaults to 64 * region server handlers configured (See the config 'hbase.regionserver.handler.count'). The math is such that by default we consider 2 MB as the result cell block size per read result and each handler will be handling a read. For 2 MB size, we need 32 buffers each of size 64 KB (See default buffer size in pool). So per handler 32 ByteBuffers(BB). We allocate twice this size as the max BBs count such that one handler can be creating the response and handing it to the RPC Responder thread and then handling a new request creating a new response cell block (using pooled buffers). Even if the responder could not send back the first TCP reply immediately, our count should allow that we should still have enough buffers in our pool without having to make temporary buffers on the heap. Again for smaller sized random row reads, tune this max count. There are lazily created buffers and the count is the max count to be pooled.

The setting for HBASE_OFFHEAPSIZE in hbase-env.sh should consider this off heap buffer pool at the RPC side also. We need to config this max off heap size for RS as a bit higher than the sum of this max pool size and the off heap cache size. The TCP layer will also need to create direct bytebuffers for TCP communication. Also the DFS client will need some off-heap to do its workings especially if short-circuit reads are configured. Allocating an extra of 1 - 2 GB for the max direct memory size has worked in tests.

If you still see GC issues even after making E2E read path off heap, look for issues in the appropriate buffer pool. Check the below RS log with INFO level:

  "Pool already reached its max capacity : XXX and no free buffers now. Consider increasing the value for 'hbase.ipc.server.reservoir.initial.max' ?"

If you are using co processors and refer the Cells in the read results, DO NOT store reference to these Cells out of the scope of the CP hook methods. Some times the CPs need store info about the cell (Like its row key) for considering in the next CP hook call etc. For such cases, pls clone the required fields of the entire Cell as per the use cases. [ See CellUtil#cloneXXX(Cell) APIs ]

• Assignee: Anoop Sam John
• Reporter: Anoop Sam John
• Created: 27/Jun/14 17:19
• Updated: 17/Jun/22 05:12
• Resolved: 28/Jan/16 05:21

Description

Umbrella jira to make sure we can have blocks cached in offheap backed cache. In the entire read path, we can refer to this offheap buffer and avoid onheap copying.
The high level items I can identify as of now are
1. Avoid the array() call on BB in read path.. (This is there in many classes. We can handle class by class)
2. Support Buffer based getter APIs in cell. In read path we will create a new Cell with backed by BB. Will need in CellComparator, Filter (like SCVF), CPs etc.
3. Avoid KeyValue.ensureKeyValue() calls in read path - This make byte copy.
4. Remove all CP hooks (which are already deprecated) which deal with KVs. (In read path)

Will add subtasks under this.

Issue Links

• depends upon: HBASE-10800 Use CellComparator instead of KVComparator
• is related to: HBASE-10772 Use ByteRanges instead of ByteBuffers in BlockCache
• is related to: HBASE-10773 Make use of ByteRanges in HFileBlock instead of ByteBuffers
• is related to: HBASE-10801 Ensure DBE interfaces can work with Cell
• is related to: HBASE-10974 Improve DBEs read performance by avoiding byte array deep copies for key[] and value[]
• is related to: HBASE-7320 Remove KeyValue.getBuffer()
• relates to: HBASE-15525 OutOfMemory could occur when using BoundedByteBufferPool during RPC bursts
• relates to: HBASE-16626 User customized RegionScanner from 1.X is incompatible with 2.0.0's off-heap part
• relates to: HBASE-20281 [DOC] upgrade section needs an explanation of changes to Bucket Cache
• relates to: HBASE-15063 Bug in MultiByteBuf#toBytes
• relates to: HBASE-15064 BufferUnderflowException after last Cell fetched from an HFile Block served from L2 offheap cache
• relates to: HBASE-14940 Make our unsafe based ops more safe
• relates to: HBASE-15379 Fake cells created in read path not implementing SettableSequenceId
• relates to: HBASE-16609 Fake cells EmptyByteBufferedCell created in read path not implementing SettableSequenceId
• relates to: HBASE-16783 Use ByteBufferPool for the header and message during Rpc response
• relates to: HBASE-11871 Avoid usage of KeyValueUtil#ensureKeyValue
• links to: RB

Sub-Tasks

• Support KeyValueCodec to encode non KeyValue cells.: Closed
• Support DirectByteBuffer usage in HFileBlock: Closed
• HFileBlock backed by Array of ByteBuffers: Closed
• Facilitate using ByteBuffer backed Cells in the HFileReader: Closed
• Ensure Cells and its implementations work with Buffers also: Closed
• Support DirectByteBuffer usage in DataBlock Encoding area: Closed
• Avoid onheap buffer copying at RPCServer#serResponse: Closed
• Column trackers and delete trackers should deal with BBs: Closed
• Prevent block eviction under us if reads are in progress from the BBs: Closed
• Filters should work with ByteBufferedCell: Closed
• Support DBB usage in Bloom and HFileIndex area: Closed
• Support BB usage in PrefixTree: Closed
• Memstore and MemstoreScanner should work with BBs.: Closed
• Redo the hfile index length optimization so cell-based rather than serialized KV key: Closed
• Unsafe based ByteBuffer Comparator: Closed
• Create ByteBuffer backed Cell: Closed
• Add unsafe putXXXUnsafe() BB methods to ByteBufferUtils: Closed
• Change DBEs to work with new BB based cell: Closed
• Tags to work with ByteBuffer: Closed
• Add ByteBufferedCell an extension to Cell: Closed
• Remove deprecated seek/reseek methods from HFileScanner: Closed
• Deperecate Filter#filterRowKey(byte[] buffer, int offset, int length) in favor of filterRowKey(Cell firstRowCell): Closed
• Deprecate RegionObserver#postScannerFilterRow CP hook with byte[],int,int args in favor of taking Cell arg: Closed
• Change ColumnTracker and SQM to deal with Cell instead of byte[], int, int: Closed
• Delayed scanner close in KeyValueHeap and StoreScanner: Closed
• Change RegionScannerImpl to deal with Cell instead of byte[], int, int: Closed
• Create MultiByteBuffer an aggregation of ByteBuffers: Closed
• Close the scanner only after Call#setResponse: Closed
• Use BufferBackedCell in read path after HBASE-12213 and HBASE-12295: Closed
• Change ByteBuff.getXXXStrictlyForward to relative position based reads: Closed
• ByteBufferUtils#compareTo small optimization: Closed
• Bloomfilter path to work with Byte buffered cells: Closed
• Reduce garbage we create: Closed
• Short circuit last byte check in CellUtil#matchingXXX methods for ByteBufferedCells: Closed
• Create the fake keys required in the scan path to avoid copy to byte[]: Closed
• Small optimization in SingleByteBuff: Closed
• Short circuit checks in ByteBufferUtils compare/equals: Closed
• Shorten ByteBufferedCell#getXXXPositionInByteBuffer method name: Closed
• Clear HFileScannerImpl#prevBlocks in between Compaction flow: Closed
• Ensure write paths work with ByteBufferedCells in case of compaction: Closed
• Support OffheapKV write in compaction with out copying data on heap: Closed
• Tightening of the CP contract: Closed
• Unnecessary lock in ByteBufferArray: Closed
• References to previous cell in read path should be avoided: Closed
• LRUBlockCache#returnBlock should try return block to Victim Handler L2 cache.: Closed
• L1 cache caching shared memory HFile block when blocks promoted from L2 to L1: Closed
• Make sure documentation is updated for the offheap Bucket cache usage: Closed