At high load it is possible that new process could fail to map shared memory [CORE3325] #1303
Submitted by: @hvlad
Test case is very artificial. There ir a lot (>= 150) of clients (isql) which connects to db, execute some script and disconnects, then repeat all over again and again. So we have a lot of connect\disconnect concurrently. If database is shutdown, then lifetime of CS process is very short, as it responds with isc_shutdown to the caller and quits.
Some time under this load i see in firebird.log :
The text was updated successfully, but these errors were encountered: