Другие доклады
-
MMikhail Maslov Банк ВТБ (ПАО)
Multi-transactions and possible issues with them
In this presentation, we cover multi-transactions, explain when they appear and how they can be monitored and checked. We also tackle the possible issue with multi-transaction ID wraparound and the problem of file absence in pg_multixact/offsets when restoring from a backup.
-
Igor Alov Yandex.Bank
Pgpool-II in the "Master-Master" mode or Network engineer’s overview of the PostgreSQL load balances solution.
One of the main needs of high-load database projects is a right configured load balancer within the database cluster. Most of the solutions I worked with, including Pgpool-II, could not fully suit business requirements. In this article me as a network engineer with broad experience will cover tricky configurations of the Pgpool-II operating in the “Master-Master” mode, as well as consider cases in which similar solutions are indispensable.
-
Pavel Konotopov inCountry
Five shades of sharding
The importance of sharding is now colossal. The size of today's databases exceeds 100 terabytes, scaling up vertically, and adding replicas containing a complete physical copy of the database is becoming difficult, especially when computational resources are scarce. Database sharding is a way to scale horizontally by partitioning data between compute nodes that are independent of each other.
In the PostgreSQL world, there are well-known scaling tools - CitusDB, and Greenplum - as well as new generation solutions - Cockroach DB, Yugabyte DB, SPQR, and Shardman.
In this talk, we will discuss the difference between these implementations, the advantages, and disadvantages of these solutions, examine the current state of implementation of sharding in vanilla PostgreSQL, and also touch on another important topic - providing data integrity and consistency guarantees across a distributed cluster.
-
Pavel Konotopov inCountry
RTT, RTO, RPO and synchronous replication
How does network latency affect performance when using synchronous replication? If a primary node fails, how long will our cluster not be reachable? How much data can be lost? Will there be any anomalies when reading from standby? We are designing our network architecture, give us a minimum RTT value for cluster solution! Many DBAs have to face these issues when designing and implementing cluster solutions. In this short talk we will tell how to quickly estimate RTT, RTO, RPO values, with the help of which Open Source tools one can test cluster solution and experimentally verify the design parameters, which metrics values can be used as a guide in design.