Memory Management Functions
This function is supported only by cluster versions 9.1.0 and later.
pg_shared_chunk_detail(contextname char(64))
Description: Queries information about all chunks requested by the memory context in a specified shared memory.
The contextname parameter indicates the name of the memory context.
Before using this function, use the pg_shared_chunk_dump(contextname char(64)) function to print related information to a file.
Return type: record
Example:
1 2 3 4 5 6 7 8 |
SELECT * FROM pg_shared_chunk_detail('pgstat file hash table'); parent | level | file_name | line_number | chunk_size | requested_number | total_size ---------------------+-------+--------------+-------------+------------+------------------+------------ pgstat file context | 2 | dynahash.cpp | 158 | 2048 | 1 | 2048 pgstat file context | 2 | dynahash.cpp | 158 | 256 | 1 | 256 pgstat file context | 2 | dynahash.cpp | 158 | 4096 | 9 | 36864 pgstat file context | 2 | dynahash.cpp | 158 | 8192 | 4 | 32768 (4 rows) |
pv_session_chunk_detail(tid int, contextname char(64))
Description: Queries information about all chunks applied for by a memory context created by a specified thread.
The tid parameter indicates the thread ID, while the contextname parameter indicates the memory context name.
To use this function, you need to use the pv_session_chunk_dump(tid int, contextname char(64)) function to print related information to a file.
Return type: record
Example:
1 2 3 4 5 6 7 8 9 |
SELECT * FROM pv_session_chunk_detail(140178810990936, 'Timezones'); parent | level | file_name | line_number | chunk_size | requested_number | total_size ------------------+-------+---------------+-------------+------------+------------------+------------ TopMemoryContext | 1 | dynahash.cpp | 158 | 1280 | 2 | 2560 TopMemoryContext | 1 | dynahash.cpp | 158 | 160 | 1 | 160 TopMemoryContext | 1 | dynahash.cpp | 158 | 2560 | 2 | 5120 TopMemoryContext | 1 | localtime.cpp | 1965 | 128 | 1 | 128 TopMemoryContext | 1 | localtime.cpp | 1965 | 448 | 1 | 448 (5 rows) |
pg_shared_chunk_dump(contextname char(64))
Description: Prints information about all chunks applied for by the memory context in the specified shared memory as a file.
The contextname parameter indicates the name of the memory context.
Return type: Boolean
Example:
1 2 3 4 5 |
SELECT * FROM pg_shared_chunk_dump('pgstat file hash table'); pg_shared_chunk_dump ---------------------- t (1 row) |
pv_session_chunk_dump(tid int, contextname char(64))
Description: Prints information about all chunks applied for by a memory context created by a specified thread into a file.
The tid parameter indicates the thread ID, while the contextname parameter indicates the memory context name.
Return type: Boolean
Example:
1 2 3 4 5 |
SELECT * FROM pv_session_chunk_dump(140472797325280, 'Timezones'); pv_session_chunk_dump ----------------------- t (1 row) |
gs_print_memory_profiling()
Description: Prints memory collection details. This function is available only for clusters of version 9.1.1.100 or later.
Return type: Boolean
Example:
1 2 3 4 5 |
SELECT * FROM gs_print_memory_profiling(); gs_print_memory_profiling ---------------------------- t (1 row) |
gs_gsc_dbstat_info(database_id bigint DEFAULT NULL::bigint)
Description: Queries the statistics of Global SysCache on the current node, such as the SysCache cache hit information and LRU-based eviction. This function is available only for clusters of version 9.1.1.100 or later.
The parameter database_id indicate the database OID.
- When database_id is left blank, statistics about all databases are queried.
- When database_id is 0, the statistics of the shared database are queried.
- When database_id is a valid database OID, the statistics of the shared database and the database are queried.
Return type: record
Return columns:
|
Column |
Type |
Description |
|---|---|---|
|
database_id |
BIGINT |
Database OID |
|
database_name |
TEXT |
Database name |
|
tup_searches |
BIGINT |
Number of system catalog cache searches |
|
tup_hits |
BIGINT |
Number of system catalog cache hits |
|
tup_miss |
BIGINT |
Number of table scans triggered by system catalog cache misses. |
|
tup_swapout_count |
BIGINT |
Number of tuples evicted by LRU |
|
tup_count |
BIGINT |
Number of tuples in the cache |
|
tup_dead |
BIGINT |
Number of tuples in the garbage queue |
|
tup_memory |
BIGINT |
Total memory of tuples in the cache |
|
total_memory |
BIGINT |
Total memory of the Global SysCache |
|
refcount |
BIGINT |
Number of times that Global SysCache is referenced by threads. |
Example:
1 2 3 4 5 6 |
SELECT * FROM gs_gsc_dbstat_info(); database_id | database_name | tup_searches | tup_hits | tup_miss | tup_swapout_count | tup_count | tup_dead | tup_memory | total_memory | refcount -------------+---------------+--------------+----------+----------+-------------------+-----------+----------+------------+--------------+---------- 0 | | 12474 | 9876 | 14 | 0 | 16 | 0 | 12560 | 697664 | 0 17185 | postgres | 10156 | 5521 | 2223 | 0 | 1398 | 0 | 752296 | 4417856 | 11 (2 rows) |
gs_gsc_clean(database_id bigint DEFAULT NULL::bigint)
Description: Clears the Global SysCache of a specified database. This function is available only for clusters of version 9.1.1.100 or later.
The parameter database_id indicate the database OID.
- When database_id is left blank, the memory of all databases is cleared.
- When database_id is 0, the memory of the shared databases is cleared.
- When database_id is a valid database OID, the memory of both the shared databases and the specified database is cleared.
Return type: bool
Example:
1 2 3 4 5 |
SELECT * FROM gs_gsc_clean(); gs_gsc_clean -------------- t (1 row) |
Feedback
Was this page helpful?
Provide feedbackThank you very much for your feedback. We will continue working to improve the documentation.See the reply and handling status in My Cloud VOC.
For any further questions, feel free to contact us through the chatbot.
Chatbot