Ilya Eremin
Ilya Eremin
CCH_flush has dbb->dbb_ast_flags & DBB_shutdown_single check to avoid PIO_flush call when a database is restoring. But a restore attachment didn't update dbb->dbb_ast_flags while setting a shutdown mode in a database...
How to reproduce on FB3: I've created a database which can show the problem: [DB_FB_3_0.zip](https://github.com/FirebirdSQL/firebird/files/8197012/DB_FB_3_0.zip) It has a table T1 with non-unique index and a procedure update_from_log which updates records...
1. The first commit fixes this script: set autoddl off; create table test (id numeric); commit; create view v_test (id) as select id from test; commit; drop view v_test; create...
FB5 crashes after these steps: ``` connect '127.0.0.1:employee' user SYSDBA password 'masterkey'; select * from mon$database; ``` Seems like the problem appeared after the profiler was merged.
If a primary record version has a transaction number which is greater than NT and seems corrupted, its state will be treated as tra_active to avoid an attempt to fetch...
Deadlock is possible when index deletion and creation are performed in the same transaction and a deleted index is used by a computed field expression. Each worker attachment loads table's...
DROP INDEX may hang for a minute if idle worker attachments hold locks in their own metadata cache
Steps to reproduce are kind of similar to #7783 but they show a completely different problem. 1. Set parameters in firebird.conf: ``` MaxParallelWorkers = 8 ParallelWorkers = 8 ``` 2....
The fix is needed to avoid potential problems in Firebird QA.