Fulltext Index Study6：Population monitor

时间：2016-06-27 15:20:45 阅读：261 评论：0 收藏：0 [点我收藏+]

标签：

一， filter daemon host

Returns information on the current activity of the filter daemon host or hosts on the server instance.

fdhost_process_id：Windows process ID of the filter daemon host.

max_thread：Maximum number of threads in the filter daemon host.

batch_count：Number of batches that are being processed in the filter daemon host. Monitors how many batches are currently being processed in the filter daemon.

查看 filter daemon host的ProcessID，最大线程数量和当前被处理的Batch数量。

select fdhost_process_id,max_thread,batch_count
from sys.dm_fts_fdhosts

通过ProcessID，能够在Task Manager中查看该Process当前消耗的CPU和内存资源。

技术分享

二，Crawl

使用Base table对 fulltext index进行填充，称作Population 或 Crawl。populating the fulltext index by using a process called a population (also known as a crawl).

1，在SQL Server中，关于Crawl的配置共有3个，使用sys.sp_configure 配置，使用sys.configurations查看配置信息

select *
from sys.configurations
where name like N‘%crawl%‘

技术分享

2，在fulltext index crawl时，查看当前消耗的memory

sys.dm_fts_memory_pools

Returns information about the shared memory pools available to the Full-Text Gatherer component for a full-text crawl or a full-text crawl range.

buffer_size:Size of each allocated buffer in the memory pool. 单位是Byte

buffer_count：Current number of shared memory buffers in the memory pool.

min_buffer_limit 和 max_buffer_limit ：Minimum/Maximum number of buffers allowed in the memory pool. 可以通过 sp_configure来修改配置。

sys.dm_fts_memory_buffers

Returns information about memory buffers belonging to a specific memory pool that are used as part of a full-text crawl or a full-text crawl range.

--returns the total shared memory owned by the Microsoft Full-Text Gatherer component  
SELECT SUM(buffer_size * buffer_count)/1024 AS total_memory_KB 
FROM sys.dm_fts_memory_pools

--return buffer details
select mp.pool_id,
    mp.buffer_size/1024 as buffer_size_KB,
    mp.buffer_count,
    mp.max_buffer_limit,
    mb.is_free,
    mb.row_count,
    mb.bytes_used,
    mb.percent_used
from sys.dm_fts_memory_pools mp
left join sys.dm_fts_memory_buffers mb
    on mp.pool_id=mb.pool_id

3，Population flow

通过sys.dm_fts_index_population 查看当前正在运行的crawl，每一次crawl都会分多个Ranges并行crawl，每一个Range可以分多个Batch进行。

population_type ：Type of population. One of the following:

1 = Full population
2 = Incremental timestamp-based population
3 = Manual update of tracked changes
4 = Background update of tracked changes.

range_count：Number of sub-ranges into which this population has been parallelized.

outstanding_batch_count：Current number of outstanding batches for this population. “outstanding” 翻译为尚待解决的，未解决的。

status：Status of this Population. Note: some of the states are transient. One of the following:

3 = Starting
5 = Processing normally
7 = Has stopped processing
11 = Population aborted
12 = Processing a semantic similarity extraction

Fulltext Index的Population，range和 batch通过Memory Address 进行Join，例句如下

SELECT ip.database_id,
    object_name(ip.table_id,ip.database_id) as TableName,
    c.name as CatalogName,
    ip.population_type,ip.population_type_description,
    ip.is_clustered_index_scan,
    ip.range_count,
    ip.completed_range_count,
    ip.outstanding_batch_count,
    ip.status,
    ip.status_description,

    pr.session_id as Range_SessionID,
    pr.processed_row_count,
    ob.batch_id    
from sys.dm_fts_index_population ip
left join sys.fulltext_catalogs c
    on ip.catalog_id=c.fulltext_catalog_id
left join sys.dm_fts_population_ranges pr
    on ip.memory_address=pr.parent_memory_address
left join sys.dm_fts_outstanding_batches ob 
    on pr.memory_address=ob.crawl_memory_address
order by ob.batch_id