Commit Graph

10 Commits

Author SHA1 Message Date
d1676c3c3d Check file descriptor number is larger than 65536 upon start (#1819) 2019-09-19 12:48:36 +08:00
981e0feb99 Check rowset is useful atomicly (#1750)
* Check rowset is useful atomicly

* Only release rowset id when it is added to unused rowset

* remove release rowset id when save rowset meta
2019-09-06 17:21:42 +08:00
a63989cc61 Use RowsetFactory to create and init RowsetWriter (#1740) 2019-09-04 17:02:43 +08:00
6f4feca3dc Add rowset id generator to FE and BE (#1678) 2019-09-02 18:51:31 +08:00
7e981b2b14 Limit the disk usage to avoid running out of disk capacity (#1702)
Set high watermark and flood stage of disk used capacity.
And forbid some operations if disk usage is too high.
2019-08-27 22:18:17 +08:00
2b2bc82ae2 Add timeout on snapshot of data (#1672)
Release snapshot when finishing or cancelling backup/restore job.
Snapshot may takes a lot disk space if not releasing them in time.
2019-08-21 21:18:53 +08:00
851b2ca3bd Remove unused code in StorageEngine (#1671) 2019-08-20 10:50:07 +08:00
dcb75729db Change cumulative compaction for decoupling storage from compution (#1576)
1. Calculate cumulative point when loading tablet first time.
2. Simplify pick rowsets logic upon delete predicate.
3. Saving meta and modify rowsets only once after cumulative compaction.
2019-08-13 18:25:56 +08:00
d938f9a6ea Implement the initial version of BetaRowset (#1568) 2019-08-06 10:40:16 +08:00
0d48a3961c Refactor Storage Engine (#1478)
NOTE: This patch would modify all Backend's data.
And this will cause a very long time to restart be.
So if you want to interferer your product environment,
you should upgrade backend one by one.

1. Refactoring be is to clarify the structure the codes.
2. Use unique id to indicate a rowset.
   Nameing rowset with tablet_id and version will lead to
   many conflicts among compaction, clone, restore.
3. Extract an rowset interface to encapsulate rowsets
   with different format.
2019-07-15 21:18:22 +08:00