2.1.7，workload group硬限生效，但是还是会超过设定的cpu_hard_limit?

Question

doris版本为2.1.7，都是按照官网设置的步骤，一个be节点下设置3个workload group，同时目前normal组也已经将cpu_hard_limit修改为10%，但是设置的3个workload group节点加起来cpu还是会超过设定值的总和？
具体运行代码如下：

I20250114 15:00:51.770592 80544 pipeline_x_fragment_context.cpp:207] PipelineXFragmentContext::prepare|query_id=d31ea08010e24092-8762121e0204ac78|fragment_id=4|pthread_id=139742613673728
I20250114 15:00:51.770875 79555 tablet.cpp:3167] calc segment delete bitmap, tablet: 14630067 rowset: 02000000015aec4ffd441523d30d0196654c16e040c4beba seg_id: 0 dummy_version: 13737 rows: 42 conflict rows: 0 bimap num: 16 cost: 11261(us)
I20250114 15:00:51.770973 79555 beta_rowset_writer.cpp:192] [Memtable Flush] construct delete bitmap tablet: 14630067, rowset_ids: 16, cur max_version: 13736, transaction_id: 16510133, cost: 11337(us), total rows: 42

I20250114 15:00:51.771513 80276 rowset_builder.cpp:287] submit calc delete bitmap task to executor, tablet_id: 14630067, txn_id: 16510133
I20250114 15:00:51.771593 80276 tablet.cpp:3186] skip to construct delete bitmap tablet: 14630067 rowset: 02000000015aec4ffd441523d30d0196654c16e040c4beba
I20250114 15:00:51.771605 80276 tablet.cpp:3538] [Before Commit] construct delete bitmap tablet: 14630067, rowset_ids to add: 0, rowset_ids to del: 0, cur max_version: 13736, transaction_id: 16510133, total rows: 42

I20250114 15:00:51.772423 80418 fragment_mgr.cpp:778] query_id: 5ae0054f14eb44aa-b056683bba55ddc0, coord_addr: TNetworkAddress(hostname=10.10.148.40, port=9020), total fragment num on current host: 33, fe process uuid: 1736524836953, query type: SELECT, report audit fe:TNetworkAddress(hostname=10.10.148.40, port=9020)
I20250114 15:00:51.772644 80418 fragment_mgr.cpp:819] Query/load id: 5ae0054f14eb44aa-b056683bba55ddc0, use workload group: TG[id = 18582232, name = wg_gmes_stop_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=21, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], is pipeline: 1
I20250114 15:00:51.772698 80418 fragment_mgr.cpp:830] Register query/load memory tracker, query/load id: 5ae0054f14eb44aa-b056683bba55ddc0 limit: 0
I20250114 15:00:51.772735 80418 pipeline_x_fragment_context.cpp:207] PipelineXFragmentContext::prepare|query_id=5ae0054f14eb44aa-b056683bba55ddc0|fragment_id=3|pthread_id=139743671154432

I20250114 15:00:51.773495 79560 tablet.cpp:3167] calc segment delete bitmap, tablet: 14630072 rowset: 02000000015aec62fd441523d30d0196654c16e040c4beba seg_id: 0 dummy_version: 13737 rows: 44 conflict rows: 0 bimap num: 15 cost: 13928(us)
I20250114 15:00:51.773495 79040 fragment_mgr.cpp:730] Removing query e738c1a99a3245ca-892e6b74caa55ac6 instance e738c1a99a3245ca-892e6b74caa55b06, all done? false
I20250114 15:00:51.773556 79560 beta_rowset_writer.cpp:192] [Memtable Flush] construct delete bitmap tablet: 14630072, rowset_ids: 15, cur max_version: 13736, transaction_id: 16510133, cost: 13973(us), total rows: 44
I20250114 15:00:51.773602 79557 tablet.cpp:3167] calc segment delete bitmap, tablet: 14630077 rowset: 02000000015aec5efd441523d30d0196654c16e040c4beba seg_id: 0 dummy_version: 13737 rows: 33 conflict rows: 0 bimap num: 15 cost: 9592(us)
I20250114 15:00:51.773656 79557 beta_rowset_writer.cpp:192] [Memtable Flush] construct delete bitmap tablet: 14630077, rowset_ids: 15, cur max_version: 13736, transaction_id: 16510133, cost: 9635(us), total rows: 33
I20250114 15:00:51.773710 85811 fragment_mgr.cpp:730] Removing query 2048b678bb964edd-b4cbc1cab8793dc0 instance 2048b678bb964edd-b4cbc1cab8793de1, all done? true
I20250114 15:00:51.773789 85811 fragment_mgr.cpp:736] Query 2048b678bb964edd-b4cbc1cab8793dc0 finished
I20250114 15:00:51.773831 78589 fragment_mgr.cpp:730] Removing query e738c1a99a3245ca-892e6b74caa55ac6 instance e738c1a99a3245ca-892e6b74caa55b17, all done? false
I20250114 15:00:51.773874 78589 fragment_mgr.cpp:730] Removing query e738c1a99a3245ca-892e6b74caa55ac6 instance e738c1a99a3245ca-892e6b74caa55b18, all done? false
I20250114 15:00:51.773917 78589 fragment_mgr.cpp:730] Removing query e738c1a99a3245ca-892e6b74caa55ac6 instance e738c1a99a3245ca-892e6b74caa55b19, all done? false
I20250114 15:00:51.773933 78589 fragment_mgr.cpp:730] Removing query e738c1a99a3245ca-892e6b74caa55ac6 instance e738c1a99a3245ca-892e6b74caa55b1a, all done? false
I20250114 15:00:51.773957 79560 vertical_segment_writer.cpp:711] add a single block 30
I20250114 15:00:51.773963 78589 fragment_mgr.cpp:730] Removing query e738c1a99a3245ca-892e6b74caa55ac6 instance e738c1a99a3245ca-892e6b74caa55b1b, all done? false
I20250114 15:00:51.773993 80276 rowset_builder.cpp:287] submit calc delete bitmap task to executor, tablet_id: 14630072, txn_id: 16510133
I20250114 15:00:51.773993 78589 fragment_mgr.cpp:730] Removing query e738c1a99a3245ca-892e6b74caa55ac6 instance e738c1a99a3245ca-892e6b74caa55b1c, all done? false
I20250114 15:00:51.774036 78589 fragment_mgr.cpp:730] Removing query e738c1a99a3245ca-892e6b74caa55ac6 instance e738c1a99a3245ca-892e6b74caa55b1d, all done? false
I20250114 15:00:51.774055 80276 tablet.cpp:3186] skip to construct delete bitmap tablet: 14630072 rowset: 02000000015aec62fd441523d30d0196654c16e040c4beba
I20250114 15:00:51.774065 79557 vertical_segment_writer.cpp:711] add a single block 33
I20250114 15:00:51.774055 78589 fragment_mgr.cpp:730] Removing query e738c1a99a3245ca-892e6b74caa55ac6 instance e738c1a99a3245ca-892e6b74caa55b1e, all done? false
I20250114 15:00:51.774070 80276 tablet.cpp:3538] [Before Commit] construct delete bitmap tablet: 14630072, rowset_ids to add: 0, rowset_ids to del: 0, cur max_version: 13736, transaction_id: 16510133, total rows: 44

I20250114 15:00:51.774314 80276 rowset_builder.cpp:287] submit calc delete bitmap task to executor, tablet_id: 14630077, txn_id: 16510133
I20250114 15:00:51.774349 80276 tablet.cpp:3186] skip to construct delete bitmap tablet: 14630077 rowset: 02000000015aec5efd441523d30d0196654c16e040c4beba
I20250114 15:00:51.774335 85843 fragment_mgr.cpp:730] Removing query 2dc96d6078d74d8e-926596819402d529 instance 2dc96d6078d74d8e-926596819402d52a, all done? false
I20250114 15:00:51.774365 80276 tablet.cpp:3538] [Before Commit] construct delete bitmap tablet: 14630077, rowset_ids to add: 0, rowset_ids to del: 0, cur max_version: 13736, transaction_id: 16510133, total rows: 33

I20250114 15:00:51.774714 78572 fragment_mgr.cpp:736] Query e738c1a99a3245ca-892e6b74caa55ac6 finished
I20250114 15:00:51.774811 80418 pipeline_x_fragment_context.cpp:207] PipelineXFragmentContext::prepare|query_id=5ae0054f14eb44aa-b056683bba55ddc0|fragment_id=2|pthread_id=139743671154432

I20250114 15:00:51.776089 85844 fragment_mgr.cpp:736] Query 2dc96d6078d74d8e-926596819402d529 finished
I20250114 15:00:51.776731 85809 fragment_mgr.cpp:730] Removing query 74bffc14234d408e-b81192b3e215c353 instance 74bffc14234d408e-b81192b3e215c374, all done? true
I20250114 15:00:51.776772 85809 fragment_mgr.cpp:736] Query 74bffc14234d408e-b81192b3e215c353 finished
I20250114 15:00:51.777033 80544 pipeline_x_fragment_context.cpp:207] PipelineXFragmentContext::prepare|query_id=d31ea08010e24092-8762121e0204ac78|fragment_id=2|pthread_id=139742613673728
I20250114 15:00:51.777720 79559 tablet.cpp:3167] calc segment delete bitmap, tablet: 14630082 rowset: 02000000015aec6bfd441523d30d0196654c16e040c4beba seg_id: 0 dummy_version: 13737 rows: 25 conflict rows: 0 bimap num: 15 cost: 9382(us)
I20250114 15:00:51.777796 79559 beta_rowset_writer.cpp:192] [Memtable Flush] construct delete bitmap tablet: 14630082, rowset_ids: 15, cur max_version: 13736, transaction_id: 16510133, cost: 9439(us), total rows: 25
I20250114 15:00:51.778304 80276 rowset_builder.cpp:287] submit calc delete bitmap task to executor, tablet_id: 14630082, txn_id: 16510133
I20250114 15:00:51.778347 79559 vertical_segment_writer.cpp:711] add a single block 35
I20250114 15:00:51.778353 80276 tablet.cpp:3186] skip to construct delete bitmap tablet: 14630082 rowset: 02000000015aec6bfd441523d30d0196654c16e040c4beba
I20250114 15:00:51.778373 80276 tablet.cpp:3538] [Before Commit] construct delete bitmap tablet: 14630082, rowset_ids to add: 0, rowset_ids to del: 0, cur max_version: 13736, transaction_id: 16510133, total rows: 25

I20250114 15:00:51.780937 85824 query_context.cpp:156] Query 2048b678bb964edd-b4cbc1cab8793dc0 deconstructed, , deregister query/load memory tracker, queryId=2048b678bb964edd-b4cbc1cab8793dc0, Limit=2.00 GB, CurrUsed=65.12 KB, PeakUsed=2.43 MB
I20250114 15:00:51.781016 80418 pipeline_x_fragment_context.cpp:207] PipelineXFragmentContext::prepare|query_id=5ae0054f14eb44aa-b056683bba55ddc0|fragment_id=1|pthread_id=139743671154432
I20250114 15:00:51.781072 85824 query_context.cpp:188] Query 2048b678bb964edd-b4cbc1cab8793dc0 deconstructed, , deregister query/load memory tracker, queryId=2048b678bb964edd-b4cbc1cab8793dc0, Limit=2.00 GB, CurrUsed=65.12 KB, PeakUsed=2.43 MB
I20250114 15:00:51.781610 80776 fragment_mgr.cpp:778] query_id: 7d06b627c6ad4236-a6b45a05fa6fbe83, coord_addr: TNetworkAddress(hostname=10.10.148.40, port=9020), total fragment num on current host: 17, fe process uuid: 1736524836953, query type: SELECT, report audit fe:TNetworkAddress(hostname=10.10.148.40, port=9020)
I20250114 15:00:51.781755 80776 fragment_mgr.cpp:819] Query/load id: 7d06b627c6ad4236-a6b45a05fa6fbe83, use workload group: TG[id = 18578947, name = wg_gmes_summary_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=34, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], is pipeline: 1
I20250114 15:00:51.781800 80776 fragment_mgr.cpp:830] Register query/load memory tracker, query/load id: 7d06b627c6ad4236-a6b45a05fa6fbe83 limit: 0
I20250114 15:00:51.781833 80776 pipeline_x_fragment_context.cpp:207] PipelineXFragmentContext::prepare|query_id=7d06b627c6ad4236-a6b45a05fa6fbe83|fragment_id=1|pthread_id=139740654507776
I20250114 15:00:51.781858 85854 query_context.cpp:156] Query 74bffc14234d408e-b81192b3e215c353 deconstructed, , deregister query/load memory tracker, queryId=74bffc14234d408e-b81192b3e215c353, Limit=2.00 GB, CurrUsed=65.12 KB, PeakUsed=2.43 MB
I20250114 15:00:51.781978 85854 query_context.cpp:188] Query 74bffc14234d408e-b81192b3e215c353 deconstructed, , deregister query/load memory tracker, queryId=74bffc14234d408e-b81192b3e215c353, Limit=2.00 GB, CurrUsed=65.12 KB, PeakUsed=2.43 MB
I20250114 15:00:51.782303 85870 fragment_mgr.cpp:730] Removing query eeac417474ef4463-8999f65faae66256 instance eeac417474ef4463-8999f65faae66277, all done? true
I20250114 15:00:51.782333 85870 fragment_mgr.cpp:736] Query eeac417474ef4463-8999f65faae66256 finished
I20250114 15:00:51.782410 80776 pipeline_x_fragment_context.cpp:207] PipelineXFragmentContext::prepare|query_id=7d06b627c6ad4236-a6b45a05fa6fbe83|fragment_id=0|pthread_id=139740654507776
I20250114 15:00:51.783603 80544 pipeline_x_fragment_context.cpp:207] PipelineXFragmentContext::prepare|query_id=d31ea08010e24092-8762121e0204ac78|fragment_id=0|pthread_id=139742613673728
I20250114 15:00:51.783639 80609 fragment_mgr.cpp:778] query_id: 4fa5bfae08c24d4a-b0b10a1d09478c22, coord_addr: TNetworkAddress(hostname=10.10.148.40, port=9020), total fragment num on current host: 33, fe process uuid: 1736524836953, query type: SELECT, report audit fe:TNetworkAddress(hostname=10.10.148.40, port=9020)
I20250114 15:00:51.783812 80609 fragment_mgr.cpp:819] Query/load id: 4fa5bfae08c24d4a-b0b10a1d09478c22, use workload group: TG[id = 18582232, name = wg_gmes_stop_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=20, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], is pipeline: 1
I20250114 15:00:51.783855 80609 fragment_mgr.cpp:830] Register query/load memory tracker, query/load id: 4fa5bfae08c24d4a-b0b10a1d09478c22 limit: 0
I20250114 15:00:51.783880 80609 pipeline_x_fragment_context.cpp:207] PipelineXFragmentContext::prepare|query_id=4fa5bfae08c24d4a-b0b10a1d09478c22|fragment_id=3|pthread_id=139742064477952
I20250114 15:00:51.784286 80609 pipeline_x_fragment_context.cpp:207] PipelineXFragmentContext::prepare|query_id=4fa5bfae08c24d4a-b0b10a1d09478c22|fragment_id=2|pthread_id=139742064477952
I20250114 15:00:51.784540 85843 query_context.cpp:156] Query 2dc96d6078d74d8e-926596819402d529 deconstructed, , deregister query/load memory tracker, queryId=2dc96d6078d74d8e-926596819402d529, Limit=2.00 GB, CurrUsed=65.12 KB, PeakUsed=2.43 MB
I20250114 15:00:51.784765 85843 query_context.cpp:188] Query 2dc96d6078d74d8e-926596819402d529 deconstructed, , deregister query/load memory tracker, queryId=2dc96d6078d74d8e-926596819402d529, Limit=2.00 GB, CurrUsed=65.12 KB, PeakUsed=2.43 MB
I20250114 15:00:51.785539 79558 tablet.cpp:3167] calc segment delete bitmap, tablet: 14630087 rowset: 02000000015aec6afd441523d30d0196654c16e040c4beba seg_id: 0 dummy_version: 13737 rows: 37 conflict rows: 0 bimap num: 16 cost: 10182(us)
I20250114 15:00:51.785594 79558 beta_rowset_writer.cpp:192] [Memtable Flush] construct delete bitmap tablet: 14630087, rowset_ids: 16, cur max_version: 13736, transaction_id: 16510133, cost: 10222(us), total rows: 37
I20250114 15:00:51.786024 79558 vertical_segment_writer.cpp:711] add a single block 45
I20250114 15:00:51.786087 80276 rowset_builder.cpp:287] submit calc delete bitmap task to executor, tablet_id: 14630087, txn_id: 16510133
I20250114 15:00:51.786154 80276 tablet.cpp:3186] skip to construct delete bitmap tablet: 14630087 rowset: 02000000015aec6afd441523d30d0196654c16e040c4beba
I20250114 15:00:51.786168 80276 tablet.cpp:3538] [Before Commit] construct delete bitmap tablet: 14630087, rowset_ids to add: 0, rowset_ids to del: 0, cur max_version: 13736, transaction_id: 16510133, total rows: 37
I20250114 15:00:51.786897 85850 query_context.cpp:156] Query eeac417474ef4463-8999f65faae66256 deconstructed, , deregister query/load memory tracker, queryId=eeac417474ef4463-8999f65faae66256, Limit=2.00 GB, CurrUsed=65.12 KB, PeakUsed=2.43 MB
I20250114 15:00:51.786991 85850 query_context.cpp:188] Query eeac417474ef4463-8999f65faae66256 deconstructed, , deregister query/load memory tracker, queryId=eeac417474ef4463-8999f65faae66256, Limit=2.00 GB, CurrUsed=65.12 KB, PeakUsed=2.43 MB
I20250114 15:00:51.787433 85841 query_context.cpp:156] Query ee0ee4cb3a0140a8-93756cfb4d5334fe deconstructed, , deregister query/load memory tracker, queryId=ee0ee4cb3a0140a8-93756cfb4d5334fe, Limit=2.00 GB, CurrUsed=65.12 KB, PeakUsed=2.43 MB
I20250114 15:00:51.787561 85841 query_context.cpp:188] Query ee0ee4cb3a0140a8-93756cfb4d5334fe deconstructed, , deregister query/load memory tracker, queryId=ee0ee4cb3a0140a8-93756cfb4d5334fe, Limit=2.00 GB, CurrUsed=65.12 KB, PeakUsed=2.43 MB

I20250114 15:00:51.789038 80758 fragment_mgr.cpp:778] query_id: a78e00a9f8a946a5-9b69e0343b7c907d, coord_addr: TNetworkAddress(hostname=10.10.148.40, port=9020), total fragment num on current host: 33, fe process uuid: 1736524836953, query type: SELECT, report audit fe:TNetworkAddress(hostname=10.10.148.40, port=9020)
I20250114 15:00:51.789266 80758 fragment_mgr.cpp:819] Query/load id: a78e00a9f8a946a5-9b69e0343b7c907d, use workload group: TG[id = 18582232, name = wg_gmes_stop_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=18, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], is pipeline: 1
I20250114 15:00:51.789305 80758 fragment_mgr.cpp:830] Register query/load memory tracker, query/load id: a78e00a9f8a946a5-9b69e0343b7c907d limit: 0
I20250114 15:00:51.789335 79040 query_context.cpp:156] Query e738c1a99a3245ca-892e6b74caa55ac6 deconstructed, , deregister query/load memory tracker, queryId=e738c1a99a3245ca-892e6b74caa55ac6, Limit=2.00 GB, CurrUsed=482.19 KB, PeakUsed=41.44 MB
I20250114 15:00:51.789348 85876 fragment_mgr.cpp:730] Removing query b369cf6cc183428f-a091230f9cde00f4 instance b369cf6cc183428f-a091230f9cde00f5, all done? false
I20250114 15:00:51.789362 80609 pipeline_x_fragment_context.cpp:207] PipelineXFragmentContext::prepare|query_id=4fa5bfae08c24d4a-b0b10a1d09478c22|fragment_id=1|pthread_id=139742064477952
I20250114 15:00:51.789338 80758 pipeline_x_fragment_context.cpp:207] PipelineXFragmentContext::prepare|query_id=a78e00a9f8a946a5-9b69e0343b7c907d|fragment_id=3|pthread_id=139740805576448

I20250114 15:00:51.789553 79040 query_context.cpp:188] Query e738c1a99a3245ca-892e6b74caa55ac6 deconstructed, , deregister query/load memory tracker, queryId=e738c1a99a3245ca-892e6b74caa55ac6, Limit=2.00 GB, CurrUsed=482.19 KB, PeakUsed=41.44 MB

I20250114 15:00:51.789850 79556 tablet.cpp:3167] calc segment delete bitmap, tablet: 14630092 rowset: 02000000015aec5bfd441523d30d0196654c16e040c4beba seg_id: 0 dummy_version: 13737 rows: 37 conflict rows: 0 bimap num: 16 cost: 16409(us)

I20250114 15:00:51.789928 79556 beta_rowset_writer.cpp:192] [Memtable Flush] construct delete bitmap tablet: 14630092, rowset_ids: 16, cur max_version: 13736, transaction_id: 16510133, cost: 16453(us), total rows: 37

I20250114 15:00:51.790221 85825 query_context.cpp:156] Query 37c18fe12d249f4-b620bdf5cfdf94f1 deconstructed, , deregister query/load memory tracker, queryId=37c18fe12d249f4-b620bdf5cfdf94f1, Limit=2.00 GB, CurrUsed=65.12 KB, PeakUsed=2.44 MB
I20250114 15:00:51.790257 85783 fragment_mgr.cpp:730] Removing query b369cf6cc183428f-a091230f9cde00f4 instance b369cf6cc183428f-a091230f9cde0114, all done? false
I20250114 15:00:51.790335 85825 query_context.cpp:188] Query 37c18fe12d249f4-b620bdf5cfdf94f1 deconstructed, , deregister query/load memory tracker, queryId=37c18fe12d249f4-b620bdf5cfdf94f1, Limit=2.00 GB, CurrUsed=65.12 KB, PeakUsed=2.44 MB
I20250114 15:00:51.790351 80276 rowset_builder.cpp:287] submit calc delete bitmap task to executor, tablet_id: 14630092, txn_id: 16510133
I20250114 15:00:51.790365 79556 vertical_segment_writer.cpp:711] add a single block 38
I20250114 15:00:51.790410 80276 tablet.cpp:3186] skip to construct delete bitmap tablet: 14630092 rowset: 02000000015aec5bfd441523d30d0196654c16e040c4beba
I20250114 15:00:51.790419 80276 tablet.cpp:3538] [Before Commit] construct delete bitmap tablet: 14630092, rowset_ids to add: 0, rowset_ids to del: 0, cur max_version: 13736,

wangbo · Accepted Answer

1 目前系统表workload_group_resource_usage可以看到各个workload group的cpu用量，他统计的cpu时间包括scan算子耗时，执行层算子（agg）耗时和导入的memtable下刷的cpu，在2.1.7版本暂时只限制了scan和执行层算子的cpu，导入的memtable下刷的暂时没有控制，预计在2.1.8版本支持。
所以可以先确认多出的那部分的cpu是不是导入的memtable引起的，可以建一个测试的workload group，只提交select语句，看下系统表的cpu使用是否和workload group的cpu配置接近

黄海军@SelectDB · Answer

跟进中，有进展会更新回帖

wangbo · Answer

topic_publish_wg，麻烦在be.info里拿下这个关键字的日志，应该会有好几行

happyday_zhang · Answer

: I20250114 14:35:10.353150 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 1, name = normal, cpu_share = 1024, memory_limit = 90.59 GB, enable_memory_overcommit = true, version = 23, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=5, spill_high_watermark=80, is_shutdown=false, query_num=0, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]
	行     5967: I20250114 14:35:10.353221 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 18578947, name = wg_gmes_summary_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=2, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]
	行     5968: I20250114 14:35:10.353248 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 18582232, name = wg_gmes_stop_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=10, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]
	行     5969: I20250114 14:35:10.353272 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 18594382, name = wg_gmes_online02_summary_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=0, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]
	行     5970: I20250114 14:35:10.353435 18799 workload_group_manager.cpp:133] [topic_publish_wg]finish clear unused workload group, time cost: 0 ms, deleted group size:0, before wg size=4, after wg size=4
	行    83993: I20250114 14:35:40.356012 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 1, name = normal, cpu_share = 1024, memory_limit = 90.59 GB, enable_memory_overcommit = true, version = 23, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=5, spill_high_watermark=80, is_shutdown=false, query_num=0, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]
	行    83994: I20250114 14:35:40.356097 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 18578947, name = wg_gmes_summary_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=33, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]
	行    83995: I20250114 14:35:40.356125 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 18582232, name = wg_gmes_stop_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=10, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]
	行    83996: I20250114 14:35:40.356146 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 18594382, name = wg_gmes_online02_summary_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=0, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]
	行    83997: I20250114 14:35:40.356310 18799 workload_group_manager.cpp:133] [topic_publish_wg]finish clear unused workload group, time cost: 0 ms, deleted group size:0, before wg size=4, after wg size=4
	行   170165: I20250114 14:36:10.357923 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 1, name = normal, cpu_share = 1024, memory_limit = 90.59 GB, enable_memory_overcommit = true, version = 23, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=5, spill_high_watermark=80, is_shutdown=false, query_num=0, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]
	行   170166: I20250114 14:36:10.357983 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 18578947, name = wg_gmes_summary_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=37, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]
	行   170167: I20250114 14:36:10.358009 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 18582232, name = wg_gmes_stop_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=10, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]
	行   170168: I20250114 14:36:10.358063 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 18594382, name = wg_gmes_online02_summary_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=0, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]
	行   170169: I20250114 14:36:10.358249 18799 workload_group_manager.cpp:133] [topic_publish_wg]finish clear unused workload group, time cost: 0 ms, deleted group size:0, before wg size=4, after wg size=4
	行   287322: I20250114 14:36:40.360513 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 1, name = normal, cpu_share = 1024, memory_limit = 90.59 GB, enable_memory_overcommit = true, version = 23, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=5, spill_high_watermark=80, is_shutdown=false, query_num=0, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]
	行   287323: I20250114 14:36:40.360545 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 18578947, name = wg_gmes_summary_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=36, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]
	行   287324: I20250114 14:36:40.360591 18799 workload_group_listener.cpp:65] [topic_publish_wg]update workload group finish, wg info=TG[id = 18582232, name = wg_gmes_stop_it, cpu_share = 1024, memory_limit = 135.89 GB, enable_memory_overcommit = true, version = 3, cpu_hard_limit = 10, scan_thread_num = 192, max_remote_scan_thread_num = 960, min_remote_scan_thread_num = 8, spill_low_watermark=50, spill_high_watermark=80, is_shutdown=false, query_num=10, read_bytes_per_second=-1, remote_read_bytes_per_second=-1], enable_cpu_hard_limit=true, cgroup cpu_shares=1024, cgroup cpu_hard_limit=10, cgroup home path=/sys/fs/cgroup/cpu/doris, list size=4, thread info=[exec num:96, real_num:96, min_num:96, max_num:96],[l_scan num:192, real_num:192, min_num:192, max_num192],[r_scan num:8, real_num:8, min_num:8, max_num:960],[mem_tab_flush num:6, real_num:6, min_num:6, max_num:6]

2.1.7，workload group硬限生效，但是还是会超过设定的cpu_hard_limit?

4 Answers