Bug #20007 DataMemory leak on drop table with var sized attributes

Submitted: 22 May 2006 19:18 Modified: 28 Jun 2006 4:19
Reporter: Tomas Ulin Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S1 (Critical)
Version:5.1.10 OS:
Assigned to: Jonas Oreland CPU Architecture:Any

[22 May 2006 19:18] Tomas Ulin
Description:
running create/drop in loop with var sized attributes causes growth in DataMemory usage

after drop, all dump 1000 still shows DataMemory used up

restart -i on a data node shows memory reclaimed with all dump 1000

How to repeat:
e.g.

ndb/ndbcluster --initial --small --diskless
./mysql-test-run.pl --ndb-connectstring=localhost:1186 --skip-master-binlog --start-and-exit ndb_basic
perl ../../mysql-test-extra-5.1/mysql-test/test_tools/scripts/stress/ndb/diskdata/cid_ndb_dd.pl -spath=var/tmp/master.sock --socket

do all dump 1000 every now and then...  it will show growth

stop script

drop table TESTER.t1

memory still used all dump 1000

ndb_mgm> 1 restart -i

all dump 1000 memory reclaimed

edit cid_ndb_dd.pl VARCHAR -> CHAR
ndb_mgm> all restart -i

run script

problem goes away...

Suggested fix:
varpages probably not released on drop table
[21 Jun 2006 11:40] Bugs System
A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/8010
[27 Jun 2006 22:02] Tomas Ulin
pushed to 5.1.12
[28 Jun 2006 4:19] Jon Stephens
Documented bugfix in 5.1.12 changelog. Closed.

Portions of this website are copyright © 2001, 2002 The PHP Group

Page generated in 0.038 sec. using MySQL 8.4.7-u6-cloud

Timestamp references displayed by the system are UTC.
Content reproduced on this site is the property of the respective copyright holders.
It is not reviewed in advance by Oracle and does not necessarily represent the opinion of Oracle or any other party.