…, I start to get nervous around 100MBps. By default Galera uses unicast TCP for replication. Because synchronous replication needs to replicate to all nodes… we add more nodes when we are using multicast. Other multicast tips We can also also bootstrap nodes using the mcast address: #wsrep… IST and SST will still use TCP unicast, so we still want to make sure those are configured to use the regular IP of…
Comment: Replication in MySQL 5.6: GTIDs benefits and limitations - Part 1
May 27, 2013 By 8 Comments
… MySQL 5.6 to MariaDB 10.0 or vice versa. You just won’t be able to use the GTID features to automatically pick the correct binlog position when switching to a new master. Old-style… compatibility is good. If you have any suggestion of how to make MariaDB compatible without merging all of the crap that…
Comment: Replication in MySQL 5.6: GTIDs benefits and limitations - Part 1
June 10, 2013 By 8 Comments
It may not be easy to simply switch from regular MySQL replication to using replication using GTIDs 1. Temporary tables – CREATE and DROP TEMPORARY…. Updates involving non-transactional storage engines – It is not possible to mix non-transactional tables (such as MYISAM) with innodb tables… transaction. The same issue can arise if slave host is using different storage engine for the respective table on master databases…
Post: Migrating between MySQL schemas with Percona Xtrabackup
June 10, 2013 By 2 Comments
… redo-logs to the dirty tablespace and voila, you have a consistent binary backup at a point in time. When running Percona Server, you can utilize the –export flag during the –apply-logs phase and then re-import those files to…_old < /tmp/orig.schema.sql Ensure innodb_import_table_from_xtrabackup = 1 (dynamic variable) Prepare the backup using the –export flag: innobackupex…
Post: The small improvements of MySQL 5.6: Duplicate Index Detection
May 31, 2013 By Leave a Comment
…the checks will be stricter: in the default SQL mode, a duplicate index will throw a warning…of all applications out there that use MySQL, for which we at Percona tend to find redundant indexes. And … key. This is the output of our tool when run on the same table, correctly identifying the redundancy:…
Post: Percona Server 5.1.69-14.7 now available: A drop in replacement for MySQL
June 10, 2013 By 3 Comments
…-Server-shared-compat package naming issue that could lead to unresolved package dependencies when installing Percona Server 5.1. Bug fixed #893860… situation when the oldest untracked log records are overwritten by the new log data. In some corner cases this could lead to… the package dependency. Bug fixed #1003776. XtraDB changed page tracking used to hold the log system mutex for the log reads needlessly…
Comment: MySQL and Percona Server in LinkBench benchmark
June 9, 2013 By 18 Comments
…. 32G BP was faster than 50GB BP even I used O_DIRECT when I tested) might not be from PCI-e flash… it might be from LinkBench. The LinkBench (at least I used at the time) seems strongly depend on GC of java… throughput. I think LinkBench should have less dependency for GC to test on fast storage, reusing objects/sessions/stmt more or…
Post: Summertime Percona MySQL training update
June 6, 2013 By 1 Comment
…, Massachusetts, USA Begins Monday, July 22, 2013 , 9am – 5pm Moving to MySQL 5.6 – Belfast, IE Begins Thursday, August 1, 2013… for use when ordering, register early and save even more as the 10% discount can be applied to the early registration price. Just use discount code mpb10 when checking to receive the discount.
Comment: ZFS on Linux and MySQL
May 31, 2013 By 17 Comments
… limit the ARC to 4GB? Also wondering if any one is using ZFS and MySQL to thin clone databases meaning using a snapshot of… meaning using a single set of datafiles that support multiple copies of the source database. Would be interested if someone wanted to… that the source is not seen for example aggregating Oracle, SQL Server and MySQL into one source that the user or…
Comment: Replication in MySQL 5.6: GTIDs benefits and limitations - Part 2
May 30, 2013 By 6 Comments
@Kristian, Well, the (slave) server *knows* whether it is replicating in parallel or not, so it doesn’t have to always manage the general case. When it knows it uses single thread replication the “last executed” assumption is safe.
