pingcap/tidb
Voir sur GitHubadmin show ddl jobs output confusing with multiple jobs
Open
#23 494 ouverte le 24 mars 2021
help wantedsig/sql-infratype/feature-request
Métriques du dépôt
- Stars
- (40 090 stars)
- Métriques de merge PR
- (Merge moyen 14j 4h) (346 PRs mergées en 30 j)
Description
Feature Request
Is your feature request related to a problem? Please describe:
Describe the feature you'd like:
I have 10 tables all trying to add an index concurrently (via sysbench):
mysql> show processlist;
+------+------+-----------+--------+---------+------+------------+--------------------------------+
| Id | User | Host | db | Command | Time | State | Info |
+------+------+-----------+--------+---------+------+------------+--------------------------------+
| 17 | root | localhost | sbtest | Query | 672 | autocommit | CREATE INDEX k_9 ON sbtest9(k) |
| 19 | root | localhost | sbtest | Query | 671 | autocommit | CREATE INDEX k_2 ON sbtest2(k) |
| 15 | root | localhost | sbtest | Query | 672 | autocommit | CREATE INDEX k_8 ON sbtest8(k) |
| 23 | root | localhost | sbtest | Query | 672 | autocommit | CREATE INDEX k_4 ON sbtest4(k) |
| 39 | root | localhost | NULL | Query | 0 | autocommit | show processlist |
+------+------+-----------+--------+---------+------+------------+--------------------------------+
5 rows in set (0.00 sec)
mysql> admin show ddl jobs;
+--------+---------+------------+--------------+----------------------+-----------+----------+-----------+---------------------+---------------------+---------+
| JOB_ID | DB_NAME | TABLE_NAME | JOB_TYPE | SCHEMA_STATE | SCHEMA_ID | TABLE_ID | ROW_COUNT | START_TIME | END_TIME | STATE |
+--------+---------+------------+--------------+----------------------+-----------+----------+-----------+---------------------+---------------------+---------+
| 79 | sbtest | sbtest9 | add index | write reorganization | 51 | 64 | 591104 | 2021-03-23 19:44:28 | NULL | running |
| 80 | sbtest | sbtest4 | add index | none | 51 | 62 | 0 | 2021-03-23 19:44:28 | NULL | none |
| 81 | sbtest | sbtest8 | add index | none | 51 | 61 | 0 | 2021-03-23 19:44:29 | NULL | none |
| 82 | sbtest | sbtest2 | add index | none | 51 | 58 | 0 | 2021-03-23 19:44:29 | NULL | none |
| 78 | sbtest | sbtest3 | add index | public | 51 | 57 | 1000001 | 2021-03-23 19:44:28 | 2021-03-23 19:55:14 | synced |
| 77 | sbtest | sbtest7 | add index | public | 51 | 60 | 1000001 | 2021-03-23 19:44:28 | 2021-03-23 19:53:28 | synced |
| 76 | sbtest | sbtest6 | add index | public | 51 | 56 | 1000001 | 2021-03-23 19:44:28 | 2021-03-23 19:51:40 | synced |
| 75 | sbtest | sbtest5 | add index | public | 51 | 59 | 1000001 | 2021-03-23 19:44:28 | 2021-03-23 19:49:54 | synced |
| 74 | sbtest | sbtest1 | add index | public | 51 | 55 | 1000001 | 2021-03-23 19:44:26 | 2021-03-23 19:48:33 | synced |
| 73 | sbtest | sbtest10 | add index | public | 51 | 53 | 1000001 | 2021-03-23 19:44:24 | 2021-03-23 19:46:29 | synced |
| 72 | sbtest | sbtest9 | create table | public | 51 | 64 | 0 | 2021-03-23 19:42:46 | 2021-03-23 19:42:47 | synced |
| 71 | sbtest | sbtest4 | create table | public | 51 | 62 | 0 | 2021-03-23 19:42:46 | 2021-03-23 19:42:47 | synced |
| 70 | sbtest | sbtest8 | create table | public | 51 | 61 | 0 | 2021-03-23 19:42:46 | 2021-03-23 19:42:47 | synced |
| 69 | sbtest | sbtest7 | create table | public | 51 | 60 | 0 | 2021-03-23 19:42:46 | 2021-03-23 19:42:47 | synced |
+--------+---------+------------+--------------+----------------------+-----------+----------+-----------+---------------------+---------------------+---------+
14 rows in set (0.00 sec)
There are two confusing parts to this output:
START_TIMEis the time the job was first created, not the time it actually started (because of queueing).- The
STATEofnonefor the queued jobs would be easier to understand if it said "queueing" or "blocked".
Describe alternatives you've considered:
These are not technically bugs. They should in theory be improved when concurrent DDL is possible, but even with concurrent I expect there will be some maximum concurrency and hence improving still helps.
Teachability, Documentation, Adoption, Migration Strategy:
N/A