Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
fix 'import cycle' for package tidb/ddl.
create sysvar TiDBFastDDL (#3) * Change TiDBFastDDL to Global scope. * TiDBFastDDL Variable usage 1, At beginning of doReorgWorkForCreateIndex, use global variable TiDBFastDDL to detemine whether enable lightning backfill process. After set up lightning, then set reorgInfo.IsLightningEnabled to true to show the lightning add index is uded for this DDL job. 2, Later, check reorgInfo.IsLightningEnabled to control the DDL job level lightning's process. Complete lightning env initlize procdeure (#4) 1. Get max open file limition for lightning, default 1024 2. Set light sorted path: 2.1 If sysVar DataDir is not a start with / data dir path, then set path to /tmp/lightning 2.2 otherwise set path to DataDir + "lightning" Check whether the lightning sorted path is exist and a dir, if yes then keep use it. otherwise create new one. 3, Set the lightning dir quota, default 10G Add one unit test uniform lightning related errmessage text to lightning_error file. (#5) complete memory track module work. refacter memmory manager to resource manager for expand to control concurrent base available cpu core number. Finsih integrate with lightning concurrency process logic. Add log for lightning processing. refactor some code` complete memory track module work. (#7) complete user cancel and exception part logic implement * complete memory track module work. * refacter memmory manager to resource manager for expand to control concurrent base available cpu core number. Finsih integrate with lightning concurrency process logic. * Add log for lightning processing. refactor some code` restore logic first stage adjust metric of ddl process progress value for lightning solution refine import cycle set up disk quota refine code and add some ut and ft. Add config paramemter TiDBlightningSortPath for setting sort parth for add index. fix mem reclaim problems Add log infromation combine the optimizes: 1, prundecode 2, json expression Add some log information
- Loading branch information