Issue Details (XML | Word | Printable)

Key: CORE-6269
Type: Bug Bug
Status: Closed Closed
Resolution: Fixed
Priority: Major Major
Assignee: Alexander Peshkov
Reporter: Alexey Kovyazin
Votes: 0
Watchers: 1

If you were logged in you would be able to see more operations.
Firebird Core

Server (on Linux) ignores KeepAlive settings, when client connection is lost, the related server connection/process is not killed

Created: 25/Mar/20 01:56 PM   Updated: 04/Apr/20 07:38 AM
Component/s: Build Issues / Porting, Installation
Affects Version/s: 3.0.5
Fix Version/s: 3.0.6

Environment: Linux centos 7, classic server

QA Status: Cannot be tested

 Description  « Hide
In a case of abnormal client's disconnect (crash), the server process is not killed, keeping database connection open

 All   Comments   Change History   Subversion Commits      Sort Order: Ascending order - Click to sort in descending order
Alexander Peshkov added a comment - 25/Mar/20 02:03 PM
That should happen on any linux using systemd (instead init + xinetd). systemd (unlike xinetd) does not provide full control of TCP/IP socket to activated by socket daemon - instead in controls that socket itself. Therefore keep alive should be enabled in it's configuration.

PS. No need to frontport to master - we do not start firebird on socket activation since FB4.

Carlos H. Cantu added a comment - 25/Mar/20 02:25 PM
Any chances of having something like this happening on Windows too? From time to time I find "zombie" connections on Windows servers too, with FB 3.

Alexander Peshkov added a comment - 25/Mar/20 02:31 PM
Definitely linux-specific issue - there is no xinetd or systemd on windows :-)

Carlos H. Cantu added a comment - 25/Mar/20 03:00 PM
Sure, but I was wondering if there is some unknown bug (in Firebird) contributing for the occurrences of zombies on Windows too.