DOCSP-34743 7.0 new concurrent read/write behavior (#7046)

* DOCSP-34743 adding dynamic concurrency info * DOCSP-34743 adding dynamic concurrency info * DOCSP-34743 adding dynamic concurrency info * DOCSP-34743 fixing prefix * DOCSP-34743 simplifying wording * DOCSP-34743 adding overload info * DOCSP-34743 adding overload info * DOCSP-34743 adding overload info
mongodb · Apr 17, 2024 · a137db9 · a137db9
1 parent 547c7d3
commit a137db9
Show file tree

Hide file tree

Showing 4 changed files with 45 additions and 5 deletions.
diff --git a/source/core/inmemory.txt b/source/core/inmemory.txt
@@ -65,8 +65,13 @@ encryption at rest configuration.
 
 .. _inmemory-concurrency:
 
-Concurrency
------------
+Transaction (Read and Write) Concurrency
+----------------------------------------
+
+.. include:: /includes/fact-dynamic-concurrency.rst
+
+Document Level Concurrency
+--------------------------
 
 The in-memory storage engine uses *document-level* concurrency control for write
 operations. As a result, multiple clients can modify different

diff --git a/source/core/wiredtiger.txt b/source/core/wiredtiger.txt
@@ -60,6 +60,22 @@ engine:
   WiredTiger doesn't allocate cache on a per-database or per-collection
   level.
 
+Transaction (Read and Write) Concurrency
+----------------------------------------
+
+.. include:: /includes/fact-dynamic-concurrency.rst
+
+To view the number of concurrent read transactions (read tickets) and
+write transactions (write tickets) allowed in the WiredTiger storage
+engine, use the :dbcommand:`serverStatus` command and see the
+:serverstatus:`wiredTiger.concurrentTransactions` parameter.
+
+.. note::
+
+   A low value of :serverstatus:`wiredTiger.concurrentTransactions` does
+   not indicate a cluster overload. Use the number of queued read and
+   write tickets as an indication of cluster overload.
+
 Document Level Concurrency
 --------------------------
 

diff --git a/source/includes/7.0-concurrent-transactions.rst b/source/includes/7.0-concurrent-transactions.rst
@@ -1,6 +1,7 @@
-Starting in MongoDB 7.0, a default algorithm is used to dynamically adjust 
-the maximum number of concurrent storage engine transactions (including both
-read and write tickets) to optimize database throughput during overload.
+Starting in version 7.0, MongoDB uses a default algorithm to dynamically
+adjust the maximum number of concurrent storage engine transactions
+(including both read and write tickets) to optimize database throughput
+during overload.
 
 The following table summarizes how to identify overload scenarios for MongoDB 
 7.0 and prior releases:

diff --git a/source/includes/fact-dynamic-concurrency.rst b/source/includes/fact-dynamic-concurrency.rst
@@ -0,0 +1,18 @@
+Starting in version 7.0, MongoDB uses a default algorithm to dynamically
+adjust the maximum number of concurrent storage engine transactions
+(read and write tickets). The dynamic concurrent storage engine
+transaction algirithm optimizes database throughput during cluster 
+overload. The maximum number of concurrent storage engine transactions
+(read and write tickets) never exceeds 128 read tickets and 128
+write tickets and may differ across nodes in a cluster. The maximum
+number of read tickets and write tickets within a single node are always
+equal. 
+
+To specify a maximum number of read and write transactions (read and
+write tickets) that the dynamic maximum can not exceed, use
+:parameter:`storageEngineConcurrentReadTransactions` and
+:parameter:`storageEngineConcurrentWriteTransactions`.
+
+If you want to disable the dynamic concurrent storage engine
+transactions algorithm, file a support request to work with a MongoDB
+Technical Services Engineer.