线上一业务sql查询在tikv 下查询55秒，在强制走tiflash下6s,但研发说目前才10天的测试数据，等后续一年数据全部进来那不是很慢？请问如何优化tiflash，提供tiflash查询速度？
1、tiflash 2个节点，副本数也是2个
2、sql 语句如下
SELECT
*
FROM
(SELECT
channel.reg_date,
iap.iap_1w_amount,
iap.iap_1_amount,
cost.cost,
channel.retention_cnt,
channel.iap_amount,
iap.iap_2w_user,
retention.retention_2w_cnt,
iap.iap_3_user,
iap.iap_1w_user,
channel.reg_cnt,
channel.iap_cnt,
retention.retention_3_cnt,
channel.install_cnt,
retention.retention_1_cnt,
iap.iap_1_user,
retention.retention_1w_cnt,
cost.show_cnt,
channel.char_cnt,
cost.click_cnt,
retention.retention_2_cnt,
iap.iap_2_user,
iap.iap_3_amount,
iap.iap_2w_amount,
iap.iap_2_amount
FROM
(SELECT
acs.reg_date,
SUM(retention_cnt) AS retention_cnt,
SUM(iap_amount) AS iap_amount,
SUM(reg_cnt) AS reg_cnt,
SUM(iap_cnt) AS iap_cnt,
SUM(install_cnt) AS install_cnt,
SUM(char_cnt) AS char_cnt
FROM
ad_channel_stat_1051 acs
LEFT JOIN ad_campaign_config c
ON acs.campaign_id = c.campaign_id
WHERE acs.user_type = 1
AND acs.reg_date >= ‘2022-11-01’
AND acs.reg_date <= ‘2022-11-30’
AND acs.gw_id = 105102
GROUP BY acs.reg_date) channel
LEFT JOIN
(SELECT
ars.reg_date,
SUM(retention_2w_cnt) AS retention_2w_cnt,
SUM(retention_3_cnt) AS retention_3_cnt,
SUM(retention_1_cnt) AS retention_1_cnt,
SUM(retention_1w_cnt) AS retention_1w_cnt,
SUM(retention_2_cnt) AS retention_2_cnt
FROM
ad_retention_stat_1051 ars
LEFT JOIN ad_campaign_config c
ON ars.campaign_id = c.campaign_id
WHERE ars.user_type = 1
AND ars.reg_date >= ‘2022-11-01’
AND ars.reg_date <= ‘2022-11-30’
AND ars.gw_id = 105102
GROUP BY ars.reg_date) retention
ON channel.reg_date = retention.reg_date
LEFT JOIN
(SELECT
ais.reg_date,
SUM(iap_1w_amount) AS iap_1w_amount,
SUM(iap_1_amount) AS iap_1_amount,
SUM(iap_2w_user) AS iap_2w_user,
SUM(iap_3_user) AS iap_3_user,
SUM(iap_1w_user) AS iap_1w_user,
SUM(iap_1_user) AS iap_1_user,
SUM(iap_2_user) AS iap_2_user,
SUM(iap_3_amount) AS iap_3_amount,
SUM(iap_2w_amount) AS iap_2w_amount,
SUM(iap_2_amount) AS iap_2_amount
FROM
ad_iap_stat_1051 ais
LEFT JOIN ad_campaign_config c
ON ais.campaign_id = c.campaign_id
WHERE ais.user_type = 1
AND ais.reg_date >= ‘2022-11-01’
AND ais.reg_date <= ‘2022-11-30’
AND ais.gw_id = 105102
GROUP BY ais.reg_date) iap
ON channel.reg_date = iap.reg_date
LEFT JOIN
(SELECT
acs.report_date,
SUM(cost) AS cost,
SUM(show_cnt) AS show_cnt,
SUM(click_cnt) AS click_cnt
FROM
ad_cost_stat_1051 acs
LEFT JOIN ad_campaign_config c
ON acs.campaign_id = c.campaign_id
WHERE acs.report_date >= ‘2022-11-01’
AND acs.report_date <= ‘2022-11-30’
AND acs.gw_id = 105102
GROUP BY acs.report_date) cost
ON channel.reg_date = cost.report_date) tmp
LIMIT 1000 OFFSET 0;
3、 执行计划如下：
************************** 1. row ***************************
id: Projection_47
estRows: 1.00
actRows: 30
task: root
access object:
execution info: time:12.6s, loops:6, Concurrency:OFF
operator info: ad_marketing_data.ad_channel_stat_1051.reg_date, Column#2313, Column#2314, Column#2350, Column#32, Column#33, Column#2315, Column#803, Column#2316, Column#2317, Column#34, Column#35, Column#804, Column#36, Column#805, Column#2318, Column#806, Column#2351, Column#37, Column#2352, Column#807, Column#2319, Column#2320, Column#2321, Column#2322
memory: 951.8 KB
disk: N/A
*************************** 2. row ***************************
id: ?..Limit_50
estRows: 1.00
actRows: 30
task: root
access object:
execution info: time:12.6s, loops:6
operator info: offset:0, count:1000
memory: N/A
disk: N/A
*************************** 3. row ***************************
id: ?..HashJoin_51
estRows: 1.00
actRows: 30
task: root
access object:
execution info: time:12.6s, loops:6, build_hash_table:{total:11.5s, fetch:11.5s, build:22.9碌s}, probe:{concurrency:5, total:1m3s, max:12.6s, probe:280.3碌s, fetch:1m3s}
operator info: left outer join, equal:[eq(ad_marketing_data.ad_channel_stat_1051.reg_date, ad_marketing_data.ad_cost_stat_1051.report_date)]
memory: 22.7 KB
disk: 0 Bytes
*************************** 4. row ***************************
id: ?..HashAgg_185(Build)
estRows: 1.00
actRows: 8
task: root
access object:
execution info: time:11.5s, loops:6, partial_worker:{wall_time:11.47633887s, concurrency:5, task_num:1, tot_wait:57.379297827s, tot_exec:353.248碌s, tot_time:57.379673406s, max:11.476236616s, p95:11.476236616s}, final_worker:{wall_time:11.476372657s, concurrency:5, task_num:5, tot_wait:57.381302115s, tot_exec:151.441碌s, tot_time:57.381462435s, max:11.476315821s, p95:11.476315821s}
operator info: group by:ad_marketing_data.ad_cost_stat_1051.report_date, funcs:sum(Column#2498)->Column#2350, funcs:sum(Column#2499)->Column#2351, funcs:sum(Column#2500)->Column#2352, funcs:firstrow(ad_marketing_data.ad_cost_stat_1051.report_date)->ad_marketing_data.ad_cost_stat_1051.report_date
memory: 109.8 KB
disk: N/A
*************************** 5. row ***************************
id: ?.?..IndexLookUp_186
estRows: 1.00
actRows: 477
task: root
access object:
execution info: time:11.5s, loops:2, index_task: {total_time: 11.3s, fetch_handle: 1.05s, build: 448.6碌s, wait: 10.2s}, table_task: {total_time: 55.5s, num: 177, concurrency: 5}
operator info:
memory: 29.9 MB
disk: N/A
*************************** 6. row ***************************
id: ?. ?..IndexRangeScan_182(Build)
estRows: 88638.98
actRows: 3545778
task: cop[tikv]
access object: table:acs, partition:p2022, index:idx_key_channel_info(report_date, media_source_id, campaign_id, adset_id, ad_id, site_id, batch_sn, region, game_id)
execution info: time:339.1ms, loops:3469, cop_task: {num: 7, max: 969.5ms, min: 334.4ms, avg: 586.1ms, p95: 969.5ms, max_proc_keys: 607349, p95_proc_keys: 607349, tot_proc: 4.02s, tot_wait: 6ms, rpc_num: 7, rpc_time: 4.1s, copr_cache_hit_ratio: 0.00}, tikv_task:{proc max:928ms, min:318ms, p80:574ms, p95:928ms, iters:3495, tasks:7}, scan_detail: {total_process_keys: 3545778, total_process_keys_size: 577770951, total_keys: 3550250, rocksdb: {delete_skipped_count: 591, key_skipped_count: 3550834, block: {cache_hit_count: 66, read_count: 5190, read_byte: 74.5 MB}}}
operator info: range:[2022-11-01,2022-11-30], keep order:false, stats:pseudo
memory: N/A
disk: N/A
*************************** 7. row ***************************
id: ?. ?..HashAgg_169(Probe)
estRows: 1.00
actRows: 477
task: cop[tikv]
access object:
execution info: time:50.7s, loops:354, cop_task: {num: 465, max: 543.5ms, min: 508碌s, avg: 166.7ms, p95: 374.9ms, max_proc_keys: 19512, p95_proc_keys: 16217, tot_proc: 1m12.7s, tot_wait: 125ms, rpc_num: 465, rpc_time: 1m17.5s, copr_cache_hit_ratio: 0.00}, tikv_task:{proc max:497ms, min:0s, p80:258ms, p95:354ms, iters:3702, tasks:465}, scan_detail: {total_process_keys: 3545778, total_process_keys_size: 723787087, total_keys: 4640582, rocksdb: {delete_skipped_count: 853981, key_skipped_count: 2010949, block: {cache_hit_count: 26394803, read_count: 10479, read_byte: 107.8 MB}}}
operator info: group by:ad_marketing_data.ad_cost_stat_1051.report_date, funcs:sum(ad_marketing_data.ad_cost_stat_1051.cost)->Column#2498, funcs:sum(ad_marketing_data.ad_cost_stat_1051.show_cnt)->Column#2499, funcs:sum(ad_marketing_data.ad_cost_stat_1051.click_cnt)->Column#2500
memory: N/A
disk: N/A
*************************** 8. row ***************************
id: ?. ?..Selection_184
estRows: 88.64
actRows: 3545778
task: cop[tikv]
access object:
execution info: tikv_task:{proc max:488ms, min:0s, p80:251ms, p95:351ms, iters:3702, tasks:465}
operator info: eq(ad_marketing_data.ad_cost_stat_1051.gw_id, 105102)
memory: N/A
disk: N/A
*************************** 9. row ***************************
id: ?. ?..TableRowIDScan_183
estRows: 88638.98
actRows: 3545778
task: cop[tikv]
access object: table:acs, partition:p2022
execution info: tikv_task:{proc max:488ms, min:0s, p80:251ms, p95:350ms, iters:3702, tasks:465}
operator info: keep order:false, stats:pseudo
memory: N/A
disk: N/A
*************************** 10. row ***************************
id: ?..Limit_55(Probe)
estRows: 1.00
actRows: 30
task: root
access object:
execution info: time:12.6s, loops:6
operator info: offset:0, count:1000
memory: N/A
disk: N/A
*************************** 11. row ***************************
id: ?..HashJoin_56
estRows: 1.00
actRows: 30
task: root
access object:
execution info: time:12.6s, loops:6, build_hash_table:{total:297.2ms, fetch:297.2ms, build:51.9碌s}, probe:{concurrency:5, total:1m3s, max:12.6s, probe:137.7碌s, fetch:1m3s}
operator info: left outer join, equal:[eq(ad_marketing_data.ad_channel_stat_1051.reg_date, ad_marketing_data.ad_iap_stat_1051.reg_date)]
memory: 71.5 KB
disk: 0 Bytes
*************************** 12. row ***************************
id: ?..HashAgg_137(Build)
estRows: 1.00
actRows: 30
task: root
access object:
execution info: time:295.6ms, loops:6, partial_worker:{wall_time:295.612215ms, concurrency:5, task_num:20, tot_wait:1.451058452s, tot_exec:22.364553ms, tot_time:1.473505767s, max:295.53882ms, p95:295.53882ms}, final_worker:{wall_time:295.682553ms, concurrency:5, task_num:25, tot_wait:1.477459291s, tot_exec:608.201碌s, tot_time:1.478077747s, max:295.645814ms, p95:295.645814ms}
operator info: group by:Column#2547, funcs:sum(Column#2536)->Column#2313, funcs:sum(Column#2537)->Column#2314, funcs:sum(Column#2538)->Column#2315, funcs:sum(Column#2539)->Column#2316, funcs:sum(Column#2540)->Column#2317, funcs:sum(Column#2541)->Column#2318, funcs:sum(Column#2542)->Column#2319, funcs:sum(Column#2543)->Column#2320, funcs:sum(Column#2544)->Column#2321, funcs:sum(Column#2545)->Column#2322, funcs:firstrow(Column#2546)->ad_marketing_data.ad_iap_stat_1051.reg_date
memory: 5.13 MB
disk: N/A
*************************** 13. row ***************************
id: ?.?..Projection_203
estRows: 0.02
actRows: 20422
task: root
access object:
execution info: time:294.4ms, loops:21, Concurrency:OFF
operator info: ad_marketing_data.ad_iap_stat_1051.iap_1w_amount, ad_marketing_data.ad_iap_stat_1051.iap_1_amount, cast(ad_marketing_data.ad_iap_stat_1051.iap_2w_user, decimal(20,0) BINARY)->Column#2538, cast(ad_marketing_data.ad_iap_stat_1051.iap_3_user, decimal(20,0) BINARY)->Column#2539, cast(ad_marketing_data.ad_iap_stat_1051.iap_1w_user, decimal(20,0) BINARY)->Column#2540, cast(ad_marketing_data.ad_iap_stat_1051.iap_1_user, decimal(20,0) BINARY)->Column#2541, cast(ad_marketing_data.ad_iap_stat_1051.iap_2_user, decimal(20,0) BINARY)->Column#2542, ad_marketing_data.ad_iap_stat_1051.iap_3_amount, ad_marketing_data.ad_iap_stat_1051.iap_2w_amount, ad_marketing_data.ad_iap_stat_1051.iap_2_amount, ad_marketing_data.ad_iap_stat_1051.reg_date, ad_marketing_data.ad_iap_stat_1051.reg_date
memory: 5.29 MB
disk: N/A
*************************** 14. row ***************************
id: ?. ?..TableReader_159
estRows: 0.02
actRows: 20422
task: root
access object:
execution info: time:284.4ms, loops:21, cop_task: {num: 185, max: 52.1ms, min: 5.17ms, avg: 21.2ms, p95: 44.7ms, rpc_num: 185, rpc_time: 3.93s, copr_cache_hit_ratio: 0.00}
operator info: data:Selection_158
memory: 119.7 KB
disk: N/A
*************************** 15. row ***************************
id: ?. ?..Selection_158
estRows: 0.02
actRows: 20422
task: cop[tiflash]
access object:
execution info: tiflash_task:{proc max:50.2ms, min:4.72ms, p80:25.4ms, p95:43ms, iters:202, tasks:185, threads:185}
operator info: eq(ad_marketing_data.ad_iap_stat_1051.gw_id, 105102), eq(ad_marketing_data.ad_iap_stat_1051.user_type, 1), ge(ad_marketing_data.ad_iap_stat_1051.reg_date, 2022-11-01 00:00:00.000000), le(ad_marketing_data.ad_iap_stat_1051.reg_date, 2022-11-30 00:00:00.000000)
memory: N/A
disk: N/A
*************************** 16. row ***************************
id: ?. ?..TableFullScan_157
estRows: 849025.00
actRows: 869752
task: cop[tiflash]
access object: table:ais, partition:p2022
execution info: tiflash_task:{proc max:50.2ms, min:3.72ms, p80:25.2ms, p95:43ms, iters:202, tasks:185, threads:185}
operator info: keep order:false, stats:pseudo
memory: N/A
disk: N/A
*************************** 17. row ***************************
id: ?..Limit_60(Probe)
estRows: 1.00
actRows: 30
task: root
access object:
execution info: time:12.6s, loops:6
operator info: offset:0, count:1000
memory: N/A
disk: N/A
*************************** 18. row ***************************
id: ?..HashJoin_61
estRows: 1.00
actRows: 30
task: root
access object:
execution info: time:12.6s, loops:6, build_hash_table:{total:676.9ms, fetch:676.8ms, build:52.8碌s}, probe:{concurrency:5, total:1m3s, max:12.6s, probe:925.4碌s, fetch:1m3s}
operator info: left outer join, equal:[eq(ad_marketing_data.ad_channel_stat_1051.reg_date, ad_marketing_data.ad_retention_stat_1051.reg_date)]
memory: 36.7 KB
disk: 0 Bytes
*************************** 19. row ***************************
id: ?..HashAgg_104(Build)
estRows: 1.00
actRows: 30
task: root
access object:
execution info: time:676.2ms, loops:6, partial_worker:{wall_time:676.151598ms, concurrency:5, task_num:1029, tot_wait:2.871705142s, tot_exec:506.818524ms, tot_time:3.379088378s, max:676.115995ms, p95:676.115995ms}, final_worker:{wall_time:676.166887ms, concurrency:5, task_num:25, tot_wait:3.380279612s, tot_exec:389.772碌s, tot_time:3.380678799s, max:676.14244ms, p95:676.14244ms}
operator info: group by:Column#2535, funcs:sum(Column#2529)->Column#803, funcs:sum(Column#2530)->Column#804, funcs:sum(Column#2531)->Column#805, funcs:sum(Column#2532)->Column#806, funcs:sum(Column#2533)->Column#807, funcs:firstrow(Column#2534)->ad_marketing_data.ad_retention_stat_1051.reg_date
memory: 2.33 MB
disk: N/A
*************************** 20. row ***************************
id: ?.?..Projection_202
estRows: 0.15
actRows: 858037
task: root
access object:
execution info: time:671.5ms, loops:1030, Concurrency:OFF
operator info: cast(ad_marketing_data.ad_retention_stat_1051.retention_2w_cnt, decimal(20,0) BINARY)->Column#2529, cast(ad_marketing_data.ad_retention_stat_1051.retention_3_cnt, decimal(20,0) BINARY)->Column#2530, cast(ad_marketing_data.ad_retention_stat_1051.retention_1_cnt, decimal(20,0) BINARY)->Column#2531, cast(ad_marketing_data.ad_retention_stat_1051.retention_1w_cnt, decimal(20,0) BINARY)->Column#2532, cast(ad_marketing_data.ad_retention_stat_1051.retention_2_cnt, decimal(20,0) BINARY)->Column#2533, ad_marketing_data.ad_retention_stat_1051.reg_date, ad_marketing_data.ad_retention_stat_1051.reg_date
memory: 275.0 KB
disk: N/A
*************************** 21. row ***************************
id: ?. ?..TableReader_126
estRows: 0.15
actRows: 858037
task: root
access object:
execution info: time:335ms, loops:1030, cop_task: {num: 608, max: 54.7ms, min: 4.67ms, avg: 15.4ms, p95: 31.8ms, rpc_num: 608, rpc_time: 9.38s, copr_cache_hit_ratio: 0.00}
operator info: data:Selection_125
memory: 1.04 MB
disk: N/A
*************************** 22. row ***************************
id: ?. ?..Selection_125
estRows: 0.15
actRows: 858037
task: cop[tiflash]
access object:
execution info: tiflash_task:{proc max:53.2ms, min:3.31ms, p80:18.4ms, p95:29.9ms, iters:693, tasks:608, threads:608}
operator info: eq(ad_marketing_data.ad_retention_stat_1051.gw_id, 105102), eq(ad_marketing_data.ad_retention_stat_1051.user_type, 1), ge(ad_marketing_data.ad_retention_stat_1051.reg_date, 2022-11-01 00:00:00.000000), le(ad_marketing_data.ad_retention_stat_1051.reg_date, 2022-11-30 00:00:00.000000)
memory: N/A
disk: N/A
*************************** 23. row ***************************
id: ?. ?..TableFullScan_124
estRows: 5987684.00
actRows: 6386801
task: cop[tiflash]
access object: table:ars, partition:p2022
execution info: tiflash_task:{proc max:53.2ms, min:2.89ms, p80:17.9ms, p95:29.8ms, iters:693, tasks:608, threads:608}
operator info: keep order:false, stats:pseudo
memory: N/A
disk: N/A
*************************** 24. row ***************************
id: ?..Limit_65(Probe)
estRows: 1.00
actRows: 30
task: root
access object:
execution info: time:12.6s, loops:6
operator info: offset:0, count:1000
memory: N/A
disk: N/A
*************************** 25. row ***************************
id: ?..HashAgg_69
estRows: 1.00
actRows: 30
task: root
access object:
execution info: time:12.6s, loops:6, partial_worker:{wall_time:12.595533488s, concurrency:5, task_num:4726, tot_wait:59.718121922s, tot_exec:3.254633373s, tot_time:1m2.975261289s, max:12.595435315s, p95:12.595435315s}, final_worker:{wall_time:12.595553229s, concurrency:5, task_num:25, tot_wait:1m2.976833995s, tot_exec:500.063碌s, tot_time:1m2.97734564s, max:12.595476252s, p95:12.595476252s}
operator info: group by:Column#2528, funcs:sum(Column#2521)->Column#32, funcs:sum(Column#2522)->Column#33, funcs:sum(Column#2523)->Column#34, funcs:sum(Column#2524)->Column#35, funcs:sum(Column#2525)->Column#36, funcs:sum(Column#2526)->Column#37, funcs:firstrow(Column#2527)->ad_marketing_data.ad_channel_stat_1051.reg_date
memory: 2.70 MB
disk: N/A
*************************** 26. row ***************************
id: ?..Projection_201
estRows: 0.16
actRows: 4838472
task: root
access object:
execution info: time:12.6s, loops:4727, Concurrency:OFF
operator info: cast(ad_marketing_data.ad_channel_stat_1051.retention_cnt, decimal(10,0) BINARY)->Column#2521, ad_marketing_data.ad_channel_stat_1051.iap_amount, cast(ad_marketing_data.ad_channel_stat_1051.reg_cnt, decimal(10,0) BINARY)->Column#2523, cast(ad_marketing_data.ad_channel_stat_1051.iap_cnt, decimal(10,0) BINARY)->Column#2524, cast(ad_marketing_data.ad_channel_stat_1051.install_cnt, decimal(10,0) BINARY)->Column#2525, cast(ad_marketing_data.ad_channel_stat_1051.char_cnt, decimal(20,0) BINARY)->Column#2526, ad_marketing_data.ad_channel_stat_1051.reg_date, ad_marketing_data.ad_channel_stat_1051.reg_date
memory: 1.32 MB
disk: N/A
*************************** 27. row ***************************
id: ?..IndexLookUp_95
estRows: 0.16
actRows: 4838472
task: root
access object:
execution info: time:10.9s, loops:4727, index_task: {total_time: 12.4s, fetch_handle: 1.13s, build: 622.1碌s, wait: 11.2s}, table_task: {total_time: 1m1s, num: 257, concurrency: 5}
operator info:
memory: 52.6 MB
disk: N/A
*************************** 28. row ***************************
id: ?..IndexRangeScan_92(Build)
estRows: 162565.67
actRows: 5191863
task: cop[tikv]
access object: table:acs, partition:p2022, index:reg_date(reg_date, media_source_id, campaign_id, adset_id, ad_id, site_id, batch_sn, region, game_id)
execution info: time:296.9ms, loops:5078, cop_task: {num: 11, max: 1.37s, min: 289.3ms, avg: 610.5ms, p95: 1.37s, max_proc_keys: 793118, p95_proc_keys: 793118, tot_proc: 6.61s, tot_wait: 1ms, rpc_num: 11, rpc_time: 6.72s, copr_cache_hit_ratio: 0.00}, tikv_task:{proc max:1.32s, min:281ms, p80:675ms, p95:1.32s, iters:5121, tasks:11}, scan_detail: {total_process_keys: 5191863, total_process_keys_size: 862314108, total_keys: 5191876, rocksdb: {delete_skipped_count: 31, key_skipped_count: 5191896, block: {cache_hit_count: 419, read_count: 7183, read_byte: 104.4 MB}}}
operator info: range:[2022-11-01,2022-11-30], keep order:false, stats:pseudo
memory: N/A
disk: N/A
*************************** 29. row ***************************
id: ?..Selection_94(Probe)
estRows: 0.16
actRows: 4838472
task: cop[tikv]
access object:
execution info: time:54s, loops:5456, cop_task: {num: 2159, max: 483.4ms, min: 361.8碌s, avg: 55.3ms, p95: 215.5ms, max_proc_keys: 18613, p95_proc_keys: 9806, tot_proc: 1m48.5s, tot_wait: 480ms, rpc_num: 2159, rpc_time: 1m59.2s, copr_cache_hit_ratio: 0.00}, tikv_task:{proc max:423ms, min:0s, p80:104ms, p95:198ms, iters:11592, tasks:2159}, scan_detail: {total_process_keys: 5191863, total_process_keys_size: 1130043877, total_keys: 5943923, rocksdb: {delete_skipped_count: 1716, key_skipped_count: 1851159, block: {cache_hit_count: 42541307, read_count: 15015, read_byte: 143.4 MB}}}
operator info: eq(ad_marketing_data.ad_channel_stat_1051.gw_id, 105102), eq(ad_marketing_data.ad_channel_stat_1051.user_type, 1)
memory: N/A
disk: N/A
*************************** 30. row ***************************
id: ?..TableRowIDScan_93
estRows: 162565.67
actRows: 5191863
task: cop[tikv]
access object: table:acs, partition:p2022
execution info: tikv_task:{proc max:423ms, min:0s, p80:104ms, p95:197ms, iters:11592, tasks:2159}
operator info: keep order:false, stats:pseudo
memory: N/A
disk: N/A
30 rows in set (12.86 sec)
5、是不是tiflash节点部署越多，走tiflash的速度就越快？
