Details
-
Bug
-
Status: Closed (View Workflow)
-
Minor
-
Resolution: Fixed
-
5.5.38-galera, 10.0.12-galera
-
None
Description
=0 is a better value that =2 because it is a background thread where durability is achieved based on isolated galera machines based on this research: openquery.com.au/blog/innodbflushlogsontrxcommit-galeracluster
upstream pull request for codership documentation: https://github.com/codership/galera/pull/86
patch:
=== modified file 'debian/additions/my.cnf'
|
--- debian/additions/my.cnf 2014-07-17 20:22:01 +0000
|
+++ debian/additions/my.cnf 2014-07-28 07:04:36 +0000
|
@@ -165,7 +165,7 @@
|
#query_cache_size=0
|
#
|
# Optional setting
|
-#innodb_flush_log_at_trx_commit=2
|
+#innodb_flush_log_at_trx_commit=0
|
|
[mysqldump]
|
quick
|
|
|
=== modified file 'support-files/rpm/server.cnf'
|
--- support-files/rpm/server.cnf 2014-07-17 20:22:01 +0000
|
+++ support-files/rpm/server.cnf 2014-07-28 07:04:52 +0000
|
@@ -25,7 +25,7 @@
|
#query_cache_size=0
|
#
|
# Optional setting
|
-#innodb_flush_log_at_trx_commit=2
|
+#innodb_flush_log_at_trx_commit=0
|
|
# this is only for embedded server
|
[embedded]
|
|
|