[ https://issues.apache.org/jira/browse/HBASE-23887?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17135181#comment-17135181 ]
Danil Lipovoy commented on HBASE-23887: --------------------------------------- [~bharathv] I found out why scan was slow - the bottle neck was net (I sent requests form other PC). Now I scan local and the net is not the problem. So I run a test scenario: 1. Scan (25 threads, batch = 100) 2. After 5 minutes add multi-gets (25 threads, batch = 100) 3. After 5 minutes switch off multi-gets (only scan again) During the test have checked the ratio between single and multi caches (added it into log file). The first run with *hbase.lru.cache.heavy.eviction.count.limit* = 10000 (disable the feature) and the second the limit = 0. !scan_and_gets.png! Take a look on the ratio (single/multi). Log file of the first run (the feature is disabled): evicted (MB): -2721, ratio 0.0, overhead (%): -1460, heavy eviction counter: 0, current caching DataBlock (%): 100 evicted (MB): -2721, ratio 0.0, overhead (%): -1460, heavy eviction counter: 0, current caching DataBlock (%): 100 evicted (MB): -2721, ratio 0.0, overhead (%): -1460, heavy eviction counter: 0, current caching DataBlock (%): 100 evicted (MB): 114, ratio 5.78, overhead (%): -43, heavy eviction counter: 0, current caching DataBlock (%): 100 < start the test evicted (MB): 5704, ratio 1.07, overhead (%): 2752, heavy eviction counter: 1, current caching DataBlock (%): 100 evicted (MB): 5245, ratio 1.06, overhead (%): 2522, heavy eviction counter: 2, current caching DataBlock (%): 100 evicted (MB): 4902, ratio 1.06, overhead (%): 2351, heavy eviction counter: 3, current caching DataBlock (%): 100 evicted (MB): 4788, ratio 1.06, overhead (%): 2294, heavy eviction counter: 4, current caching DataBlock (%): 100 evicted (MB): 5132, ratio 1.06, overhead (%): 2466, heavy eviction counter: 5, current caching DataBlock (%): 100 evicted (MB): 5018, ratio 1.07, overhead (%): 2409, heavy eviction counter: 6, current caching DataBlock (%): 100 evicted (MB): 5244, ratio 1.06, overhead (%): 2522, heavy eviction counter: 7, current caching DataBlock (%): 100 evicted (MB): 5019, ratio 1.07, overhead (%): 2409, heavy eviction counter: 8, current caching DataBlock (%): 100 evicted (MB): 4902, ratio 1.06, overhead (%): 2351, heavy eviction counter: 9, current caching DataBlock (%): 100 evicted (MB): 4904, ratio 1.06, overhead (%): 2352, heavy eviction counter: 10, current caching DataBlock (%): 100 evicted (MB): 5017, ratio 1.06, overhead (%): 2408, heavy eviction counter: 11, current caching DataBlock (%): 100 evicted (MB): 4563, ratio 1.06, overhead (%): 2181, heavy eviction counter: 12, current caching DataBlock (%): 100 evicted (MB): 4338, ratio 1.06, overhead (%): 2069, heavy eviction counter: 13, current caching DataBlock (%): 100 evicted (MB): 4789, ratio 1.06, overhead (%): 2294, heavy eviction counter: 14, current caching DataBlock (%): 100 evicted (MB): 4902, ratio 1.06, overhead (%): 2351, heavy eviction counter: 15, current caching DataBlock (%): 100 evicted (MB): 5130, ratio 1.06, overhead (%): 2465, heavy eviction counter: 16, current caching DataBlock (%): 100 evicted (MB): 5017, ratio 1.06, overhead (%): 2408, heavy eviction counter: 17, current caching DataBlock (%): 100 evicted (MB): 4795, ratio 1.06, overhead (%): 2297, heavy eviction counter: 18, current caching DataBlock (%): 100 evicted (MB): 4905, ratio 1.07, overhead (%): 2352, heavy eviction counter: 19, current caching DataBlock (%): 100 evicted (MB): 4911, ratio 1.06, overhead (%): 2355, heavy eviction counter: 20, current caching DataBlock (%): 100 evicted (MB): 5019, ratio 1.06, overhead (%): 2409, heavy eviction counter: 21, current caching DataBlock (%): 100 evicted (MB): 5134, ratio 1.06, overhead (%): 2467, heavy eviction counter: 22, current caching DataBlock (%): 100 evicted (MB): 5016, ratio 1.06, overhead (%): 2408, heavy eviction counter: 23, current caching DataBlock (%): 100 evicted (MB): 4450, ratio 1.06, overhead (%): 2125, heavy eviction counter: 24, current caching DataBlock (%): 100 evicted (MB): 4904, ratio 1.07, overhead (%): 2352, heavy eviction counter: 25, current caching DataBlock (%): 100 evicted (MB): 4561, ratio 1.06, overhead (%): 2180, heavy eviction counter: 26, current caching DataBlock (%): 100 evicted (MB): 4334, ratio 1.06, overhead (%): 2067, heavy eviction counter: 27, current caching DataBlock (%): 100 evicted (MB): 4789, ratio 1.06, overhead (%): 2294, heavy eviction counter: 28, current caching DataBlock (%): 100 evicted (MB): 4792, ratio 1.06, overhead (%): 2296, heavy eviction counter: 29, current caching DataBlock (%): 100 evicted (MB): 4903, ratio 1.07, overhead (%): 2351, heavy eviction counter: 30, current caching DataBlock (%): 100 evicted (MB): 4791, ratio 1.06, overhead (%): 2295, heavy eviction counter: 31, current caching DataBlock (%): 100 evicted (MB): 4677, ratio 1.07, overhead (%): 2238, heavy eviction counter: 32, current caching DataBlock (%): 100 evicted (MB): 4903, ratio 1.06, overhead (%): 2351, heavy eviction counter: 33, current caching DataBlock (%): 100 evicted (MB): 4791, ratio 1.06, overhead (%): 2295, heavy eviction counter: 34, current caching DataBlock (%): 100 evicted (MB): 4677, ratio 1.06, overhead (%): 2238, heavy eviction counter: 35, current caching DataBlock (%): 100 evicted (MB): 4909, ratio 1.06, overhead (%): 2354, heavy eviction counter: 36, current caching DataBlock (%): 100 evicted (MB): 4678, ratio 1.06, overhead (%): 2239, heavy eviction counter: 37, current caching DataBlock (%): 100 evicted (MB): 4788, ratio 1.06, overhead (%): 2294, heavy eviction counter: 38, current caching DataBlock (%): 100 evicted (MB): 5133, ratio 1.06, overhead (%): 2466, heavy eviction counter: 39, current caching DataBlock (%): 100 evicted (MB): 5017, ratio 1.06, overhead (%): 2408, heavy eviction counter: 40, current caching DataBlock (%): 100 evicted (MB): 4903, ratio 1.06, overhead (%): 2351, heavy eviction counter: 41, current caching DataBlock (%): 100 evicted (MB): 4789, ratio 1.07, overhead (%): 2294, heavy eviction counter: 42, current caching DataBlock (%): 100 evicted (MB): 4790, ratio 1.06, overhead (%): 2295, heavy eviction counter: 43, current caching DataBlock (%): 100 evicted (MB): 4789, ratio 1.06, overhead (%): 2294, heavy eviction counter: 44, current caching DataBlock (%): 100 evicted (MB): 4449, ratio 1.04, overhead (%): 2124, heavy eviction counter: 45, current caching DataBlock (%): 100 evicted (MB): 5672, ratio 1.11, overhead (%): 2736, heavy eviction counter: 46, current caching DataBlock (%): 100 < added gets evicted (MB): 5658, ratio 1.11, overhead (%): 2729, heavy eviction counter: 47, current caching DataBlock (%): 100 evicted (MB): 5891, ratio 1.11, overhead (%): 2845, heavy eviction counter: 48, current caching DataBlock (%): 100 evicted (MB): 6008, ratio 1.11, overhead (%): 2904, heavy eviction counter: 49, current caching DataBlock (%): 100 evicted (MB): 5832, ratio 1.15, overhead (%): 2816, heavy eviction counter: 50, current caching DataBlock (%): 100 evicted (MB): 6393, ratio 1.12, overhead (%): 3096, heavy eviction counter: 51, current caching DataBlock (%): 100 evicted (MB): 6218, ratio 1.11, overhead (%): 3009, heavy eviction counter: 52, current caching DataBlock (%): 100 evicted (MB): 5939, ratio 1.12, overhead (%): 2869, heavy eviction counter: 53, current caching DataBlock (%): 100 evicted (MB): 6369, ratio 1.11, overhead (%): 3084, heavy eviction counter: 54, current caching DataBlock (%): 100 evicted (MB): 6198, ratio 1.14, overhead (%): 2999, heavy eviction counter: 55, current caching DataBlock (%): 100 evicted (MB): 6170, ratio 1.13, overhead (%): 2985, heavy eviction counter: 56, current caching DataBlock (%): 100 evicted (MB): 6014, ratio 1.12, overhead (%): 2907, heavy eviction counter: 57, current caching DataBlock (%): 100 evicted (MB): 6359, ratio 1.11, overhead (%): 3079, heavy eviction counter: 58, current caching DataBlock (%): 100 evicted (MB): 6164, ratio 1.1, overhead (%): 2982, heavy eviction counter: 59, current caching DataBlock (%): 100 evicted (MB): 6283, ratio 1.11, overhead (%): 3041, heavy eviction counter: 60, current caching DataBlock (%): 100 evicted (MB): 6037, ratio 1.12, overhead (%): 2918, heavy eviction counter: 61, current caching DataBlock (%): 100 evicted (MB): 6216, ratio 1.11, overhead (%): 3008, heavy eviction counter: 62, current caching DataBlock (%): 100 evicted (MB): 6053, ratio 1.11, overhead (%): 2926, heavy eviction counter: 63, current caching DataBlock (%): 100 evicted (MB): 6026, ratio 1.1, overhead (%): 2913, heavy eviction counter: 64, current caching DataBlock (%): 100 evicted (MB): 5759, ratio 1.11, overhead (%): 2779, heavy eviction counter: 65, current caching DataBlock (%): 100 evicted (MB): 6313, ratio 1.1, overhead (%): 3056, heavy eviction counter: 66, current caching DataBlock (%): 100 evicted (MB): 5979, ratio 1.11, overhead (%): 2889, heavy eviction counter: 67, current caching DataBlock (%): 100 evicted (MB): 6237, ratio 1.12, overhead (%): 3018, heavy eviction counter: 68, current caching DataBlock (%): 100 evicted (MB): 6354, ratio 1.13, overhead (%): 3077, heavy eviction counter: 69, current caching DataBlock (%): 100 evicted (MB): 6409, ratio 1.11, overhead (%): 3104, heavy eviction counter: 70, current caching DataBlock (%): 100 evicted (MB): 5788, ratio 1.11, overhead (%): 2794, heavy eviction counter: 71, current caching DataBlock (%): 100 evicted (MB): 5526, ratio 1.11, overhead (%): 2663, heavy eviction counter: 72, current caching DataBlock (%): 100 evicted (MB): 6032, ratio 1.12, overhead (%): 2916, heavy eviction counter: 73, current caching DataBlock (%): 100 evicted (MB): 6307, ratio 1.11, overhead (%): 3053, heavy eviction counter: 74, current caching DataBlock (%): 100 evicted (MB): 6277, ratio 1.1, overhead (%): 3038, heavy eviction counter: 75, current caching DataBlock (%): 100 evicted (MB): 6091, ratio 1.11, overhead (%): 2945, heavy eviction counter: 76, current caching DataBlock (%): 100 evicted (MB): 5905, ratio 1.1, overhead (%): 2852, heavy eviction counter: 77, current caching DataBlock (%): 100 evicted (MB): 6185, ratio 1.1, overhead (%): 2992, heavy eviction counter: 78, current caching DataBlock (%): 100 evicted (MB): 6169, ratio 1.12, overhead (%): 2984, heavy eviction counter: 79, current caching DataBlock (%): 100 evicted (MB): 6268, ratio 1.11, overhead (%): 3034, heavy eviction counter: 80, current caching DataBlock (%): 100 evicted (MB): 6246, ratio 1.12, overhead (%): 3023, heavy eviction counter: 81, current caching DataBlock (%): 100 evicted (MB): 6332, ratio 1.13, overhead (%): 3066, heavy eviction counter: 82, current caching DataBlock (%): 100 evicted (MB): 6001, ratio 1.11, overhead (%): 2900, heavy eviction counter: 83, current caching DataBlock (%): 100 evicted (MB): 5931, ratio 1.1, overhead (%): 2865, heavy eviction counter: 84, current caching DataBlock (%): 100 evicted (MB): 6197, ratio 1.1, overhead (%): 2998, heavy eviction counter: 85, current caching DataBlock (%): 100 evicted (MB): 6291, ratio 1.11, overhead (%): 3045, heavy eviction counter: 86, current caching DataBlock (%): 100 evicted (MB): 6283, ratio 1.11, overhead (%): 3041, heavy eviction counter: 87, current caching DataBlock (%): 100 evicted (MB): 6040, ratio 1.07, overhead (%): 2920, heavy eviction counter: 88, current caching DataBlock (%): 100 evicted (MB): 5358, ratio 1.07, overhead (%): 2579, heavy eviction counter: 89, current caching DataBlock (%): 100 < interrupt gets evicted (MB): 4903, ratio 1.06, overhead (%): 2351, heavy eviction counter: 90, current caching DataBlock (%): 100 evicted (MB): 4902, ratio 1.06, overhead (%): 2351, heavy eviction counter: 91, current caching DataBlock (%): 100 evicted (MB): 5016, ratio 1.06, overhead (%): 2408, heavy eviction counter: 92, current caching DataBlock (%): 100 evicted (MB): 4904, ratio 1.06, overhead (%): 2352, heavy eviction counter: 93, current caching DataBlock (%): 100 evicted (MB): 5019, ratio 1.07, overhead (%): 2409, heavy eviction counter: 94, current caching DataBlock (%): 100 evicted (MB): 4904, ratio 1.06, overhead (%): 2352, heavy eviction counter: 95, current caching DataBlock (%): 100 evicted (MB): 5130, ratio 1.06, overhead (%): 2465, heavy eviction counter: 96, current caching DataBlock (%): 100 evicted (MB): 4904, ratio 1.06, overhead (%): 2352, heavy eviction counter: 97, current caching DataBlock (%): 100 evicted (MB): 4903, ratio 1.07, overhead (%): 2351, heavy eviction counter: 98, current caching DataBlock (%): 100 evicted (MB): 4904, ratio 1.06, overhead (%): 2352, heavy eviction counter: 99, current caching DataBlock (%): 100 evicted (MB): 5017, ratio 1.06, overhead (%): 2408, heavy eviction counter: 100, current caching DataBlock (%): 100 evicted (MB): 4903, ratio 1.06, overhead (%): 2351, heavy eviction counter: 101, current caching DataBlock (%): 100 evicted (MB): 4788, ratio 1.07, overhead (%): 2294, heavy eviction counter: 102, current caching DataBlock (%): 100 evicted (MB): 4789, ratio 1.06, overhead (%): 2294, heavy eviction counter: 103, current caching DataBlock (%): 100 evicted (MB): 4903, ratio 1.06, overhead (%): 2351, heavy eviction counter: 104, current caching DataBlock (%): 100 evicted (MB): 4903, ratio 1.06, overhead (%): 2351, heavy eviction counter: 105, current caching DataBlock (%): 100 evicted (MB): 4674, ratio 1.06, overhead (%): 2237, heavy eviction counter: 106, current caching DataBlock (%): 100 evicted (MB): 4906, ratio 1.06, overhead (%): 2353, heavy eviction counter: 107, current caching DataBlock (%): 100 evicted (MB): 4904, ratio 1.06, overhead (%): 2352, heavy eviction counter: 108, current caching DataBlock (%): 100 evicted (MB): 5018, ratio 1.06, overhead (%): 2409, heavy eviction counter: 109, current caching DataBlock (%): 100 evicted (MB): 4789, ratio 1.06, overhead (%): 2294, heavy eviction counter: 110, current caching DataBlock (%): 100 evicted (MB): 4789, ratio 1.06, overhead (%): 2294, heavy eviction counter: 111, current caching DataBlock (%): 100 evicted (MB): 4902, ratio 1.07, overhead (%): 2351, heavy eviction counter: 112, current caching DataBlock (%): 100 evicted (MB): 5017, ratio 1.06, overhead (%): 2408, heavy eviction counter: 113, current caching DataBlock (%): 100 evicted (MB): 4674, ratio 1.06, overhead (%): 2237, heavy eviction counter: 114, current caching DataBlock (%): 100 evicted (MB): 4674, ratio 1.06, overhead (%): 2237, heavy eviction counter: 115, current caching DataBlock (%): 100 evicted (MB): 4902, ratio 1.06, overhead (%): 2351, heavy eviction counter: 116, current caching DataBlock (%): 100 evicted (MB): 4793, ratio 1.07, overhead (%): 2296, heavy eviction counter: 117, current caching DataBlock (%): 100 evicted (MB): 4790, ratio 1.06, overhead (%): 2295, heavy eviction counter: 118, current caching DataBlock (%): 100 evicted (MB): 4902, ratio 1.06, overhead (%): 2351, heavy eviction counter: 119, current caching DataBlock (%): 100 evicted (MB): 4562, ratio 1.06, overhead (%): 2181, heavy eviction counter: 120, current caching DataBlock (%): 100 evicted (MB): 4788, ratio 1.06, overhead (%): 2294, heavy eviction counter: 121, current caching DataBlock (%): 100 evicted (MB): 3801, ratio 1.02, overhead (%): 1800, heavy eviction counter: 122, current caching DataBlock (%): 100 evicted (MB): 0, ratio 0.0, overhead (%): -100, heavy eviction counter: 0, current caching DataBlock (%): 100 So the ratio about 1.05 - 1.1 all the way. continue in the next message... > BlockCache performance improve by reduce eviction rate > ------------------------------------------------------ > > Key: HBASE-23887 > URL: https://issues.apache.org/jira/browse/HBASE-23887 > Project: HBase > Issue Type: Improvement > Components: BlockCache, Performance > Reporter: Danil Lipovoy > Assignee: Danil Lipovoy > Priority: Minor > Attachments: 1582787018434_rs_metrics.jpg, > 1582801838065_rs_metrics_new.png, BC_LongRun.png, > BlockCacheEvictionProcess.gif, cmp.png, evict_BC100_vs_BC23.png, > eviction_100p.png, eviction_100p.png, eviction_100p.png, gc_100p.png, > graph.png, image-2020-06-07-08-11-11-929.png, > image-2020-06-07-08-19-00-922.png, image-2020-06-07-12-07-24-903.png, > image-2020-06-07-12-07-30-307.png, image-2020-06-08-17-38-45-159.png, > image-2020-06-08-17-38-52-579.png, image-2020-06-08-18-35-48-366.png, > read_requests_100pBC_vs_23pBC.png, requests_100p.png, requests_100p.png, > requests_new2_100p.png, requests_new_100p.png, scan.png, scan_and_gets.png, > wave.png > > > Hi! > I first time here, correct me please if something wrong. > All latest information is here: > [https://docs.google.com/document/d/1X8jVnK_3lp9ibpX6lnISf_He-6xrHZL0jQQ7hoTV0-g/edit?usp=sharing] > I want propose how to improve performance when data in HFiles much more than > BlockChache (usual story in BigData). The idea - caching only part of DATA > blocks. It is good becouse LruBlockCache starts to work and save huge amount > of GC. > Sometimes we have more data than can fit into BlockCache and it is cause a > high rate of evictions. In this case we can skip cache a block N and insted > cache the N+1th block. Anyway we would evict N block quite soon and that why > that skipping good for performance. > --- > Some information below isn't actual > --- > > > Example: > Imagine we have little cache, just can fit only 1 block and we are trying to > read 3 blocks with offsets: > 124 > 198 > 223 > Current way - we put the block 124, then put 198, evict 124, put 223, evict > 198. A lot of work (5 actions). > With the feature - last few digits evenly distributed from 0 to 99. When we > divide by modulus we got: > 124 -> 24 > 198 -> 98 > 223 -> 23 > It helps to sort them. Some part, for example below 50 (if we set > *hbase.lru.cache.data.block.percent* = 50) go into the cache. And skip > others. It means we will not try to handle the block 198 and save CPU for > other job. In the result - we put block 124, then put 223, evict 124 (3 > actions). > See the picture in attachment with test below. Requests per second is higher, > GC is lower. > > The key point of the code: > Added the parameter: *hbase.lru.cache.data.block.percent* which by default = > 100 > > But if we set it 1-99, then will work the next logic: > > > {code:java} > public void cacheBlock(BlockCacheKey cacheKey, Cacheable buf, boolean > inMemory) { > if (cacheDataBlockPercent != 100 && buf.getBlockType().isData()) > if (cacheKey.getOffset() % 100 >= cacheDataBlockPercent) > return; > ... > // the same code as usual > } > {code} > > Other parameters help to control when this logic will be enabled. It means it > will work only while heavy reading going on. > hbase.lru.cache.heavy.eviction.count.limit - set how many times have to run > eviction process that start to avoid of putting data to BlockCache > hbase.lru.cache.heavy.eviction.bytes.size.limit - set how many bytes have to > evicted each time that start to avoid of putting data to BlockCache > By default: if 10 times (100 secunds) evicted more than 10 MB (each time) > then we start to skip 50% of data blocks. > When heavy evitions process end then new logic off and will put into > BlockCache all blocks again. > > Descriptions of the test: > 4 nodes E5-2698 v4 @ 2.20GHz, 700 Gb Mem. > 4 RegionServers > 4 tables by 64 regions by 1.88 Gb data in each = 600 Gb total (only FAST_DIFF) > Total BlockCache Size = 48 Gb (8 % of data in HFiles) > Random read in 20 threads > > I am going to make Pull Request, hope it is right way to make some > contribution in this cool product. > -- This message was sent by Atlassian Jira (v8.3.4#803005)