How to interprete db_stats report for -m option

Question

Hi
please help me to interprete the db_stats report (for -m option).
What is the total number of pages created by DB? (It is "Current page count" or "The number of page allocations").
what is " Total number of times hash chains searched for a page" , "Total number of hash chain entries checked for page" and " Requested pages found in the cache" ( we see really huge numbers there when XML file size is increased)
Thanks,
Santhosh

Answer 1

Hi Santhosh,
Here are some explanations on the memory pool (cache) statistics:
- “The number of page allocations” (st_alloc)
     This accounts for the total number of page allocations made by BDB. We allocate pages in the memory pool when we create, open and rename a file in the memory pool, when we get a page from a database file, when we resize the memory pool, when we create and extend the free page list, etc. This value is grow-only.
- “Current total page count” (st_pages)
     This represents the number of pages found in the cache (both clean, with non-updated content, and dirty, with updated content). So, this equals the sum of the “Current clean page count” (st_page_clean) and the “Current dirty page count” (st_page_dirty).
- “Requested pages found in the cache” (st_cache_hit)
     This represents the total number of requests for database pages that were satisfied with pages existing in the memory pool. The percent value represents the cache hit rate, that is, how many requested pages were found in the cache, and it’s calculated using the formula “st_cache_hit / (st_cache_hit + st_cache_miss)”.
- “Requested pages not found in the cache” (st_cache_miss)
     This is the total number of requested pages that were not served directly from cache and had to be read from disk. You can determine a miss rate value, using the following formula: “st_cache_miss / (st_cache_miss + st_cache_hit)”; or, you may substract the hit rate from 100%. A higher hit rate means a hotter cache; details on how to appropriately size the cache is here:
http://www.oracle.com/technology/documentation/berkeley-db/db/ref/am_conf/cachesize.html
- “Number of hash buckets used for page location” (st_hash_buckets)
     We organize and manage the cache buffer space using hash buckets; each hash bucket has a hash chain associated where its entries (hash chain entries) refer to buffers in the memory pool space. Thus, this value represents the number of hash buckets we’re using for administering pages in the memory pool.
- “Total number of times hash chains searched for a page” (st_hash_searches)
     This value accounts for the total number of times we’ve inspected the hash chains when searching for a page in the cache (memory pool).
- “Total number of hash chain entries checked for a page” (st_hash_examined)
     This value represents the total number of entries in cache’s hash bucket chains we’ve examined when looking for a page.
When the container’s/database’s size increases, the number of pages that comprise the file will increase as well. The statistics you’ve mentioned are grow-only, so their values increase during the lifetime of the application.
Let me know if this information is helpful and if you need further clarification.
Regards,
Andrei