Issue Details (XML | Word | Printable)

Key: CORE-3809
Type: Bug Bug
Status: In Progress In Progress
Priority: Major Major
Assignee: Dmitry Yemanov
Reporter: Pavel Zotov
Votes: 1
Watchers: 5

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

Cannot shutdown the database or the whole server while it performs a longish savepoint cleanup during the rollback

Created: 07/Apr/12 02:29 PM   Updated: 09/Apr/12 05:51 AM
Component/s: Engine
Affects Version/s: 2.1.0, 2.1.1, 2.1.2, 2.1.3, 3.0 Initial, 2.5.0, 2.1.4, 2.5.1
Fix Version/s: None

 Description  « Hide
After some highload test was interrupted (250 connects were killed via pskill.exe) I could not wait till FB will complete all ROLLBACKs.
total count of threads decreased too slowly and after 10 minutes it was ~100 (i.e. the decreasing rate was only 15 threads per minute):

bash-3.2$ ps -FLC fb_smp_server|wc -l

So I decided to restart FB:

[root@fbozzy logs]# /etc/init.d/firebird stop
Stopping Firebird server [default]: [ OK ]

But it seems that processes of fbguart and fb_smp_server are alive:

[root@fbozzy logs]# ps -e|grep fb
16709 ? 00:00:00 fbguard
29969 ? 00:00:00 fb_smp_server

[root@fbozzy logs]# ls -l /var/run/firebird/
total 0

How to make them be stopped ? Only via kill ?


ISQL Version: WI-V2.5.1.26351 Firebird 2.5
Server version:
Firebird/linux AMD64 (access method), version "LI-V2.5.2.26448 Firebird 2.5"
Firebird/linux AMD64 (remote server), version "LI-V2.5.2.26448 Firebird 2.5/tcp (fbozzy)/P12"
Firebird/x86/Windows NT (remote interface), version "WI-V2.5.1.26351 Firebird 2.5/tcp (tlprg)/P12"
on disk structure version 11.2

 All   Comments   Change History   Subversion Commits      Sort Order: Ascending order - Click to sort in descending order
Alexander Peshkov added a comment - 09/Apr/12 05:24 AM
It's very interesting point. Probably be we should change rollback method when shutdown is active.