When creating a tablet, it is necessary to select a disk from all disks that
meet the requirements on the BE node to store the tablet.
In Doris, the current disk selection strategy is to randomly select a disk
from all disks that meet the requirements for tablet creation.
After the cluster has been running for a long time, we found that the
distribution of the number of tablets on different disks in a BE node is unbalanced.
In order to solve this problem, we introduced the algorithm of "two random choices"
for disk selection when creating the tablet:
(1) Select two disks from all disks that meet the requirements on the BE node randomly；
(2) Choose the disk with a smaller number of tablet from the two disks selected in (1) for tablet creation.
1. Disable the MySQL client and LZO library by default when building the Doris.
MySQL client library is used for MySQL external table feature.
This feature will be replaced by the new ODBC external table soon.
LZO library is used to compress/decompress data of some old data format of Doris,
which is no longer used anymore.
2. Add missing license to some files.
3. For all non-Apache-License code, all are explained in NOTICE file and the corresponding license is declared.
4. Remove the js source code from webroot, it will be downloaded as thirdparty
Since the Segment V2 has been released for a long time, we should make it as default storage format for newly created table.
This CL mainly changes:
1. For all newly created tables, their default storage format is Segment V2.
2. For all already exist tablets, their storage format remain unchanged.
3. Fix bugs described in Fix#4384 and Fix#4385
We have changed most of our serialization methods to json. In order to be compatible with previous data, these classes still retain the readFields method. Some prs that involve modifying metadata often modify the readFields method. To avoid this, we should Mark these methods as Deprecated #4398
* Implements the grammar of the batch delete #4051
* Process create, alter table when table has delete sign column
* Support the syntax for enabling the delete column
* Automatically filtered deleted data in the select statement.
* Automatically add delete sign when create rollup table
TODO:
* Optimize the reading and compaction logic on the be side, so that the data marked as deleted will be completely deleted during base compaction
```
SELECT *
FROM
(SELECT cs_order_number,
cs_warehouse_sk
FROM catalog_sales
WHERE cs_order_number = 125005
AND cs_warehouse_sk = 4) cs1
LEFT SEMI JOIN
(SELECT cs_order_number,
cs_warehouse_sk
FROM catalog_sales
WHERE cs_order_number = 125005) cs2
ON cs1.cs_order_number = cs2.cs_order_number
AND cs1.cs_warehouse_sk <> cs2.cs_warehouse_sk;
```
The above query has an equal predicate and a not equal predicate.
If there exists not equal preidcate, the build table should be remained
as it is. So the deduplication should be removed.
Define Expr will not serialized in Column `toThrift`.
1. When adding partition, different indexes should use their own keys type
instead of using the keys type of base table uniformly.
`
2. There are two kinds of define expr in Column , one is analyzed, and the other is not analyzed.
Currently, analyzed define expr is only used when creating materialized views, so the define expr in RollupJob must be analyzed.
In other cases, such as define expr in `MaterializedIndexMeta`, it may not be analyzed after being relayed.
When executing the load, the analyzed define expr (such as to_bitmap(cast(k1, varchar))) will not be analyzed again.
Only a cast function will be added to the inner layer(such as to_bitmap(cast(cast(k1 ,int), varchar))) which is analyzed too.
The define expr that has not been analyzed (such as cast(k1, varchar)) will be analyzed when executing the load.
1. Input the correct keys type when mv is updated.
The keys type of mv should be used in schema change job rather then keys type of base table.
Otherwise, the be will core and thrown exception "Create replicas failed".
2. Forbidden add non-key column on agg mv directly when base table is duplicate model
If a dup table has a agg mv, user will not add a non-key column on mv.
The non-key column can only be added to dup index.
The rewrite rule named `CountToSum` does not distinguish between `Count` and `Count distinct` which causes `Count distinct` is rewritten as `Sum` incorrectly.
So this commit modified matching rule.
When the function is `Count distinct`, the rewrite rule will not take effect.
Fixed#4381
replace is an user defined function, which is to replace all old substrings with a new substring in a string, as follow:
mysql> select replace("http://www.baidu.com:9090", "9090", "");
+------------------------------------------------------+
| replace('http://www.baidu.com:9090', '9090', '') |
+------------------------------------------------------+
| http://www.baidu.com: |
+------------------------------------------------------+
The original Doris bitmap aggregation function has poor performance on the intersection and union set of bitmap cardinality of more than one billion. There are two reasons for this. The first is that when the bitmap cardinality is large, if the data size exceeds 1g, the network / disk IO time consumption will increase; The second point is that all the sink data of the back-end be instance are transferred to the top node for intersection and union calculation, which leads to the pressure on the top single node and becomes the bottleneck.
My solution is to create a fixed schema table based on the Doris fragmentation rule, and hash fragment the ID range based on the bitmap, that is, cut the ID range vertically to form a small cube. Such bitmap blocks will become smaller and evenly distributed on all back-end be instances. Based on the schema table, some new high-performance udaf aggregation functions are developed. All Scan nodes participate in intersection and union calculation, and top nodes only summarize
The design goal is that the base number of bitmap is more than 10 billion, and the response time of cross union set calculation of 100 dimensional granularity is within 5 s.
There are three udaf functions in this commit: orthogonal_bitmap_intersect_count, orthogonal_bitmap_union_count, orthogonal_bitmap_intersect.
Compaction rules optimization, the detail problem description and design to see #4164.
This pr commits 2 functions:
(1) add the cumulative policy configable, and implement original policy.
(2) implement universal policy, the optimization version in #4164.
Sometimes we want to detect the hotspot of a cluster, for example, hot scanned tablet, hot wrote tablet,
but we have no insight about tablets in the cluster.
This patch introduce tablet level metrics to help to achieve this object, now support 4 metrics on tablets: `query_scan_bytes `, `query_scan_rows `, `flush_bytes `, `flush_count `.
However, one BE may holds hundreds of thousands of tablets, so I add a parameter for the metrics HTTP request,
and not return tablet level metrics by default.
After PR: #4135, If a mem tracker has parent, it should be created by 'CreateTracker'.
So I removed other unused constructors.
And also fix the bug described in #4344
When setting global variables, such as `set global default_rowset_type=beta`,
the operation is not correctly persisted.
This CL change the fe meta version to 90.
---------------
The main reason for this problem is that for the modification of global variable,
we directly use Java's reflection mechanism to modify static member variables in `GlobalVariable` class.
But in the persistence method of the `set` operation, we only persist the value stored
in the `globalSessionVariable` variable, and this variable does not contain Global Variable.
So I added a new OperationType: `OP_GLOBAL_VARIABLE_V2`,
and added a `GlobalVarPersistInfo` class to record all changes.
In some very special circumstances, such as code bugs, or human misoperation, etc.,
all replicas of some tablets may be lost. In this case, the data has been substantially lost.
However, in some scenarios, the business still hopes to ensure that the query will not
report errors even if there is data loss, and reduce the perception of the user layer.
At this point, we can use the blank Tablet to fill the missing replica to ensure that the query can be executed normally.
Add a new FE config `recover_with_empty_tablet`. default is false. true means to use empty tablet to fill the missing one.
Also fix a bug in Fix#4274
Doris use HashTable to implement except.
If user send A except B except C, first do A except B and then except C.
After A except B, HashTable will be rebuild.
There is a bug here to throw some rows.
The partition column of table also must be the key in materialized view.
If not, when user wants to add partition of table, the be will core.
The materialized view could not create partition correctly when partition column has been aggregated.
(1) Add LargeInt cast to date and datatime, see #3864
LargeInt can cast to date and datatime. Fix this error:
Unable to find _ZN5doris13CastFunctions16cast_to_date_valEPN9doris_udf15FunctionContextERKNS1_11LargeIntValE
(2) Add local timezone info to stale_version_path_json_doc rest api
Add timezone to "last create time" field.
{
"path id": "1",
"last create time": "1970-01-01 10:46:40 +0800",
"path list": "1 -> [2-3] -> [4-5]"
},
and add timezone to the test unix, see #4121 .
