Improve validation of recovery_target_xid GUC values.
Previously, the recovery_target_xid GUC values were not sufficiently validated.
As a result, clearly invalid inputs such as the string "bogus", a decimal value
like "1.1", or 0 (a transaction ID smaller than the minimum valid value of 3)
were u
Add system view pg_stat_recovery
This commit introduces pg_stat_recovery, that exposes at SQL level the
state of recovery as tracked by XLogRecoveryCtlData in shared memory,
maintained by the startup process. This new view includes the following
fields, that are useful for monitoring purposes on
Fix inconsistent elevel in pg_sync_replication_slots() retry logic.
The commit 0d2d4a0ec3 allowed pg_sync_replication_slots() to retry sync
attempts, but missed a case, when WAL prior to a slot's
confirmed_flush_lsn is not yet flushed locally.
By changing the elevel from ERROR to LOG, we allow th
Fix order of columns in pg_stat_recovery
recovery_last_xact_time is listed before current_chunk_start_time in the
documentation, the function definition and the view definition, but
their order was reversed in the code.
Thinko in 01d485b142e4. Mea culpa.
Author: Shinya Kato
Discussion:
https:
Fix inconsistency with HeapTuple freeing in extended_stats_funcs.c
heap_freetuple() is a thin wrapper doing a pfree(), and the function
import_pg_statistic(), introduced by ba97bf9cb7b4, had the idea to call
directly pfree() rather than the "dedicated" heap tuple routine.
upsert_pg_statistic_ext_
Improve tests for recovery_target_timeline GUC.
Commit fd7d7b71913 added regression tests to verify recovery_target_timeline
settings. To confirm that invalid values are rejected, those tests started the
server with an invalid setting and then verified that startup failed. While
functionally corre
Fix publisher shutdown hang caused by logical walsender busy loop.
Previously, when logical replication was running, shutting down
the publisher could cause the logical walsender to enter a busy loop
and prevent the publisher from completing shutdown.
During shutdown, the logical walsender waits
Fix publisher shutdown hang caused by logical walsender busy loop.
Previously, when logical replication was running, shutting down
the publisher could cause the logical walsender to enter a busy loop
and prevent the publisher from completing shutdown.
During shutdown, the logical walsender waits
Fix publisher shutdown hang caused by logical walsender busy loop.
Previously, when logical replication was running, shutting down
the publisher could cause the logical walsender to enter a busy loop
and prevent the publisher from completing shutdown.
During shutdown, the logical walsender waits
Fix publisher shutdown hang caused by logical walsender busy loop.
Previously, when logical replication was running, shutting down
the publisher could cause the logical walsender to enter a busy loop
and prevent the publisher from completing shutdown.
During shutdown, the logical walsender waits
Fix publisher shutdown hang caused by logical walsender busy loop.
Previously, when logical replication was running, shutting down
the publisher could cause the logical walsender to enter a busy loop
and prevent the publisher from completing shutdown.
During shutdown, the logical walsender waits
Fix publisher shutdown hang caused by logical walsender busy loop.
Previously, when logical replication was running, shutting down
the publisher could cause the logical walsender to enter a busy loop
and prevent the publisher from completing shutdown.
During shutdown, the logical walsender waits
Fix handling of updated tuples in the MERGE statement
This branch missed the IsolationUsesXactSnapshot() check. That led to EPQ on
repeatable read and serializable isolation levels. This commit fixes the
issue and provides a simple isolation check for that. Backpatch through v15
where MERGE sta
oauth: Add TLS support for oauth_validator tests
The oauth_validator tests don't currently support HTTPS, which makes
testing PGOAUTHCAFILE difficult. Add a localhost certificate to
src/test/ssl and make use of it in oauth_server.py.
In passing, explain the hardcoded use of IPv4 in our issuer ide
oauth: Report cleanup errors as warnings on stderr
Using conn->errorMessage for these "shouldn't-happen" cases will only
work if the connection itself fails. Our SSL and password callbacks
print WARNINGs when they find themselves in similar situations, so
follow their lead.
Reviewed-by: Zsolt Par
libpq: Add PQgetThreadLock() to mirror PQregisterThreadLock()
Allow libpq clients to retrieve the current pg_g_threadlock pointer with
PQgetThreadLock(). Single-threaded applications could already do this in
a convoluted way:
pgthreadlock_t tlock;
tlock = PQregisterThreadLock(NULL);
Fix handling of updated tuples in the MERGE statement
This branch missed the IsolationUsesXactSnapshot() check. That led to EPQ on
repeatable read and serializable isolation levels. This commit fixes the
issue and provides a simple isolation check for that. Backpatch through v15
where MERGE sta
Fix handling of updated tuples in the MERGE statement
This branch missed the IsolationUsesXactSnapshot() check. That led to EPQ on
repeatable read and serializable isolation levels. This commit fixes the
issue and provides a simple isolation check for that. Backpatch through v15
where MERGE sta
Add PageGetPruneXid() helper
This is similar to the other page accessors in bufpage.h. It improves
readability and avoids long lines.
Author: Melanie Plageman
Reviewed-by: Chao Li
Discussion: https://postgr.es/m/BD8B69E7-26D8-4706-9164-597C6AE57812%40gmail.com
Branch
--
master
Details
---
Exit after fatal errors in client-side compression code.
It looks like whoever wrote the astreamer (nee bbstreamer) code
thought that pg_log_error() is equivalent to elog(ERROR), but
it's not; it just prints a message. So all these places tried to
continue on after a compression or decompression
Exit after fatal errors in client-side compression code.
It looks like whoever wrote the astreamer (nee bbstreamer) code
thought that pg_log_error() is equivalent to elog(ERROR), but
it's not; it just prints a message. So all these places tried to
continue on after a compression or decompression
Exit after fatal errors in client-side compression code.
It looks like whoever wrote the astreamer (nee bbstreamer) code
thought that pg_log_error() is equivalent to elog(ERROR), but
it's not; it just prints a message. So all these places tried to
continue on after a compression or decompression
Exit after fatal errors in client-side compression code.
It looks like whoever wrote the astreamer (nee bbstreamer) code
thought that pg_log_error() is equivalent to elog(ERROR), but
it's not; it just prints a message. So all these places tried to
continue on after a compression or decompression
Exit after fatal errors in client-side compression code.
It looks like whoever wrote the astreamer (nee bbstreamer) code
thought that pg_log_error() is equivalent to elog(ERROR), but
it's not; it just prints a message. So all these places tried to
continue on after a compression or decompression
Move commonly used context into PruneState and simplify helpers
heap_page_prune_and_freeze() and many of its helpers use the heap
buffer, block number, and page. Other helpers took the heap page and
didn't use it. Initializing these values once during
prune_freeze_setup() simplifies the helpers' i
Fix handling of updated tuples in the MERGE statement
This branch missed the IsolationUsesXactSnapshot() check. That led to EPQ on
repeatable read and serializable isolation levels. This commit fixes the
issue and provides a simple isolation check for that. Backpatch through v15
where MERGE sta
Fix handling of updated tuples in the MERGE statement
This branch missed the IsolationUsesXactSnapshot() check. That led to EPQ on
repeatable read and serializable isolation levels. This commit fixes the
issue and provides a simple isolation check for that. Backpatch through v15
where MERGE sta
Prefix PruneState->all_{visible,frozen} with set_
The PruneState had members called "all_visible" and "all_frozen" which
reflect not the current state of the page but the state it could be in
once pruning and freezing have been executed. These are then saved in
the PruneFreezeResult so the caller
Simplify creation of built-in functions with non-default ACLs.
Up to now, to create such a function, one had to make a pg_proc.dat
entry and then modify it with GRANT/REVOKE commands, which we put in
system_functions.sql. That seems a little ugly though, because it
violates the idea of having a s
Be more wary of false matches in initdb's replace_token().
Do not replace the target string unless the occurrence is surrounded
by whitespace or line start/end. This avoids potential false match
to a substring of a field. While we've not had trouble with that
up to now, the next patch creates ha
Refactor code retrieving string for RecoveryPauseState
This refactoring is going to be useful in an upcoming commit, to avoid
some code duplication with the function pg_get_wal_replay_pause_state(),
that returns a string for the recovery pause state.
Refactoring opportunity noticed while hacking
31 matches
Mail list logo
