[typo](docs) Fix some ambiguous descriptions (#23912)
This commit is contained in:
@ -351,7 +351,8 @@ Only the case of a single BE is discussed here. If the user cluster has multiple
|
||||
|
||||
For example, for a 100G file, the number of BEs in the cluster is 10
|
||||
max_broker_concurrency = 10
|
||||
max_bytes_per_broker_scanner >= 10G = 100G / 10
|
||||
# >= 10G = 100G / 10
|
||||
max_bytes_per_broker_scanner = 1069547520
|
||||
````
|
||||
|
||||
After modification, all BEs will process the import task concurrently, each BE processing part of the original file.
|
||||
@ -364,7 +365,8 @@ Only the case of a single BE is discussed here. If the user cluster has multiple
|
||||
The amount of data processed by a single BE of the current import task / the slowest import speed of the user Doris cluster (MB/s) >= the timeout time of the current import task >= the amount of data processed by a single BE of the current import task / 10M/s
|
||||
|
||||
For example, for a 100G file, the number of BEs in the cluster is 10
|
||||
timeout >= 1000s = 10G / 10M/s
|
||||
# >= 1000s = 10G / 10M/s
|
||||
timeout = 1000
|
||||
````
|
||||
|
||||
3. When the user finds that the timeout time calculated in the second step exceeds the default import timeout time of 4 hours
|
||||
|
||||
@ -1159,7 +1159,7 @@ current running txns on db xxx is xx, larger than limit xx
|
||||
|
||||
#### `max_bytes_per_broker_scanner`
|
||||
|
||||
默认值:500 * 1024 * 1024 * 1024L (500G)
|
||||
默认值:`500 * 1024 * 1024 * 1024L` (500G)
|
||||
|
||||
是否可以动态配置:true
|
||||
|
||||
|
||||
@ -351,7 +351,8 @@ Broker Load 需要借助 Broker 进程访问远端存储,不同的 Broker 需
|
||||
|
||||
比如一个 100G 的文件,集群的 BE 个数为 10 个
|
||||
max_broker_concurrency = 10
|
||||
max_bytes_per_broker_scanner >= 10G = 100G / 10
|
||||
# >= 10G = 100G / 10
|
||||
max_bytes_per_broker_scanner = 1069547520
|
||||
```
|
||||
|
||||
修改后,所有的 BE 会并发的处理导入任务,每个 BE 处理原始文件的一部分。
|
||||
@ -364,7 +365,8 @@ Broker Load 需要借助 Broker 进程访问远端存储,不同的 Broker 需
|
||||
当前导入任务单个 BE 处理的数据量 / 用户 Doris 集群最慢导入速度(MB/s) >= 当前导入任务的 timeout 时间 >= 当前导入任务单个 BE 处理的数据量 / 10M/s
|
||||
|
||||
比如一个 100G 的文件,集群的 BE 个数为 10个
|
||||
timeout >= 1000s = 10G / 10M/s
|
||||
# >= 1000s = 10G / 10M/s
|
||||
timeout = 1000
|
||||
```
|
||||
|
||||
3. 当用户发现第二步计算出的 timeout 时间超过系统默认的导入最大超时时间 4小时
|
||||
|
||||
Reference in New Issue
Block a user