[ https://issues.apache.org/jira/browse/IGNITE-17735?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17607891#comment-17607891 ]
Vladimir Steshin edited comment on IGNITE-17735 at 10/28/22 5:18 AM: --------------------------------------------------------------------- Datastreamer with 'allowOverwrite==true' and PRIMARY_SYNC persistent cache may cause heap issue or consume increased heap amount with persistent caches. There is related fixing 'perNodeParallelOperations()' setting. But I met this issue with trivial research like `HeapConsumptionDataStreamerTest.src`. Default parallel batches amount for persistent caches looks too much. The setting is historically for in-memory caches. What happens: streamer keep sending more and more streamer batches to process while receiving node collects backup updates futures, requests. Similarly, backup node accumulates incoming update requests stucking at disk writes. See 'DS_heap_consumption.png' for example. Suggestion. We might bring reduced default parallel batches number for persistent caches `IgniteDataStreamer#DFLT_PARALLEL_OPS_PERSISTENT_MULTIPLIER`(PR #10343). Why sending more? Helps a lot, reduces heap utilization even if there is no OOMe. Better solution would be a backpressure. Not sure it worth the case. Did some benchmarks. For persistent caches `CPUs x 2` seems enough. was (Author: vladsz83): Datastreamer with 'allowOverwrite==true' and PRIMARY_SYNC persistent cache may cause heap issue or consume increased heap amount. There is related 'perNodeParallelOperations()' setting. What discouraged, I met this issue with trivial research like few servers, simple cache and just trying data streaming with various persistence and loading settings (like `HeapConsumptionDataStreamerTest.src`). Think user may meet the same. The default value might be adjusted for persistant caches. Streamer node may not wait for backup updates. And keep sending more and more streamer batches to process. The receiving node collects related to backup updates futures, requests. The same happens on backup node: collecting update incoming update requests stucking at disk writes. See 'DS_heap_consumption.png' for example. Suggestion: bring reduced default parallel batches number for persistent caches `IgniteDataStreamer#DFLT_PARALLEL_OPS_PERSISTENT_MULTIPLIER` (PR #10343). Did estimation benchmarks. Even in-memory benchmarks (like 'bench_inmem_isolated_pc2.txt') shows 2 or may be 4 batches per threads seems enough. For persistent caches, `CPUs x 2` seems enough. See `bench_persistent_results_Isolated_pc1.txt` and `bench_persistent_results_Individual_pc1.txt` > Datastreamer may consume heap with allowOverwtire=='true'. > ---------------------------------------------------------- > > Key: IGNITE-17735 > URL: https://issues.apache.org/jira/browse/IGNITE-17735 > Project: Ignite > Issue Type: Sub-task > Reporter: Vladimir Steshin > Assignee: Vladimir Steshin > Priority: Major > Labels: ise > Attachments: DS_heap_consumption.png, DS_heap_consumption_2.png, > HeapConsumptionDataStreamerTest.src, > bench_persistent_results_Individual_pc1.txt, > bench_persistent_results_Isolated_pc1.txt > > Time Spent: 1h > Remaining Estimate: 0h > -- This message was sent by Atlassian Jira (v8.20.10#820010)