Postgres 10 highlight - Wait mode for pg_ctl promote

2016-10-09 05:50:22 +0000

pg_ctl is a control utility on top of the postgres binary allowing to do many operations on a server. A couple of its subcommands offer a way to wait for the operation to finish before exiting pg_ctl, which is useful to ensure the state of the server before moving on with other things in the upper application layer. Sub-commands start, restart and register (Windows-only to register Postgres as a service in the in-core SCM), are not waited to finish by default, but this can be done by using the -w switch. Reversely, the sub-command stop implies to wait, and this can be disabled by using the switch -W. In Postgres 10, the sub-command promote has gained this option, per the following commit:

Postgres 10 highlight - pg_basebackup and data durability

2016-10-04 08:15:09 +0000

One of the work areas happening for Postgres 10 currently in development is the problem related to data durability when using the in-core utilities to take data dumps, base backups, or anything related to Postgres like WAL segments. In short, the goal is for users to be sure that any data taken from Postgres using the in-core tools is still present and consistent on disk even if a power failure occurs once the binary writing and flushing the data is done. A couple of days ago the following commit has landed in the code tree to improve the situation for pg_basebackup and pg_receivexlog:

Postgres 10 highlight - recovery_target_lsn

2016-09-13 01:31:43 +0000

When performing point-in-time recovery, Postgres offers a variety of ways to stop recovery, or WAL replay at a given point using different ways of estimating the stop point:

Postgres 10 highlight - ShmemAlloc and ShmemAllocNoError

2016-09-04 07:16:22 +0000

While doing some work on Postgres 10, I have noticed that in a lot of code paths some system calls related to memory allocation, like malloc(), realloc() or strdup() simply missed to check for NULL values, causing a crash. Note that crashes are highly unlikely to happen, still the result of this investigation to make the code of Postgres cleaner, with problems detected as well by Heikki Linnakangas and Aleksander Alekseev, has resulted on a set of commits, like that, that and finally that.

Postgres 9.6 - Upcoming changes for pg_get functions on invalid objects

2016-07-27 13:14:54 +0000

Today the following commit has landed on the master branch of PostgreSQL, meaning that it will be included in the upcoming 9.6:

« Prev 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 Next »

Michael Paquier - PostgreSQL committer

Postgres 10 highlight - Wait mode for pg_ctl promote

Postgres 10 highlight - pg_basebackup and data durability

Postgres 10 highlight - recovery_target_lsn

Postgres 10 highlight - ShmemAlloc and ShmemAllocNoError

Postgres 9.6 - Upcoming changes for pg_get functions on invalid objects

Search

Social

RSS Feeds

Sponsor