June 7, 2010 at 2:57 am
Hi all.
I am trying to investigate details of the i/o consumption on my server with DMF dm_io_virtual_file_stats.
The first result of num_of_bytes_read value was 4727558504448 for one of the database data file.
The second result after 15 minutes was 4757704859648.
Then I subtract first result from the second one and result is 30146355200. It is number of bytes which have been read from file in 15 minutes.
I twice divide that value by 1024 and my final result in megabyte is 28749.8.
But that is near 20 Gb.
I have verified this value with Performance monitor (Disk Read Bytes/Sec). The result was comparable 10 - 20 Gb for reading per 15 min.
How can I explain such large value?
Which addition metrics could help me?
The size of the whole database file is near 9 Gb and Cache Hit Ratio is near 99.
There are about 150 users work with database.
June 7, 2010 at 3:14 am
I guess you have to drill down some more to identify which queries that performed physical reads in the same period. You can find information about this in sys.dm_exec_query_stats. Take a snapshot of the data before and after your 15 minutes period and calculate the difference in physical reads in that period.
I'm not sure if the number in total_physical_reads is # of pages or if it is physical read operations. If it is read operations then it can be hard to determine if the reads are pages, extents or from log file data. BOL doesn't say much about that.
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply