MySQL Bugs: #39539: REPORT MemoryUsge suddenly become unusable

Bug #39539	REPORT MemoryUsge suddenly become unusable
Submitted:	19 Sep 2008 15:34	Modified:	24 Nov 2008 2:26
Reporter:	Mikiya Okuno	Email Updates:
Status:	No Feedback	Impact on me:	None
Category:	MySQL Cluster: Cluster (NDB) storage engine	Severity:	S3 (Non-critical)
Version:	6.3.16-telco	OS:	Any
Assigned to:	Assigned Account	CPU Architecture:	Any

Description:
After several inserts to a NDB table, ALL REPORT MemoryUsage become unusable and nothing's displayed. I do not know what is the trigger of the problem, and it doesn't happen so often. Once the command becomes unusable, it will not recover unless ndb_mgm reconnects to ndb_mgmd. Other commands like SHOW works well when the problem is happening.

How to repeat:
1. Create table

mysql> create table t1 (a bigint unsigned primary key, b varchar(200), c timestamp) engine=ndb;

2. Insert thousands of rows.

3. Display memory usage.

ndb_mgm> show
Cluster Configuration
---------------------
[ndbd(NDB)]	4 node(s)
id=41	@127.0.0.1  (mysql-5.1.24 ndb-6.3.16, Nodegroup: 0, Master)
id=42	@127.0.0.1  (mysql-5.1.24 ndb-6.3.16, Nodegroup: 0)
id=43	@127.0.0.1  (mysql-5.1.24 ndb-6.3.16, Nodegroup: 1)
id=44	@127.0.0.1  (mysql-5.1.24 ndb-6.3.16, Nodegroup: 1)

[ndb_mgmd(MGM)]	1 node(s)
id=1	@127.0.0.1  (mysql-5.1.24 ndb-6.3.16)

[mysqld(API)]	2 node(s)
id=51	@127.0.0.1  (mysql-5.1.24 ndb-6.3.16)
id=52	@127.0.0.1  (mysql-5.1.24 ndb-6.3.16)

ndb_mgm> all status
Node 41: started (mysql-5.1.24 ndb-6.3.16)
Node 42: started (mysql-5.1.24 ndb-6.3.16)
Node 43: started (mysql-5.1.24 ndb-6.3.16)
Node 44: started (mysql-5.1.24 ndb-6.3.16)

ndb_mgm> all report memoryusage

ndb_mgm> all report memoryusage

4. After reconnect, it works well.

ndb_mgm> quit
open-machine:~ mikiya$ /usr/local/mysql-cluster-6.3.16-telco/bin/ndb_mgm
-- NDB Cluster -- Management Client --
ndb_mgm> all report memoryusage
Connected to Management Server at: localhost:1186

ndb_mgm> Node 41: Data usage is 7%(74 32K pages of total 1024)
Node 41: Index usage is 2%(27 8K pages of total 1312)
Node 42: Data usage is 7%(74 32K pages of total 1024)
Node 42: Index usage is 2%(27 8K pages of total 1312)
Node 43: Data usage is 7%(74 32K pages of total 1024)
Node 43: Index usage is 1%(26 8K pages of total 1312)
Node 44: Data usage is 7%(74 32K pages of total 1024)
Node 44: Index usage is 1%(26 8K pages of total 1312)

Suggested fix:
N/A

Hi Mikiya,

You get the bug from binary package or source codes.
I can't reproduce it from source codes of mysql-5.1-telco-6.3 and mysql-5.1-telco-6.2 tree.

Hi Guangbao,

I see the problem on what I built from the source code.

Hi Mikiya,

But from the output information, the source codes you used isn't the newest source codes of telco-6.3.

The version of newest telco-6.3 is mysql-5.1.28 ndb-6.3.19.

ndb_mgm> show
Connected to Management Server at: localhost:9311
Cluster Configuration
---------------------
[ndbd(NDB)]     2 node(s)
id=1    @127.0.0.1  (mysql-5.1.28 ndb-6.3.19, Nodegroup: 0, Master)
id=2    @127.0.0.1  (mysql-5.1.28 ndb-6.3.19, Nodegroup: 0)

[ndb_mgmd(MGM)] 1 node(s)
id=3    @127.0.0.1  (mysql-5.1.28 ndb-6.3.19)

[mysqld(API)]   12 node(s)
id=4    @127.0.0.1  (mysql-5.1.28 ndb-6.3.19)
id=5    @127.0.0.1  (mysql-5.1.28 ndb-6.3.19)
id=6    @127.0.0.1  (mysql-5.1.28 ndb-6.3.19)
id=7    @127.0.0.1  (mysql-5.1.28 ndb-6.3.19)
id=8    @127.0.0.1  (mysql-5.1.28 ndb-6.3.19)
id=9    @127.0.0.1  (mysql-5.1.28 ndb-6.3.19)
id=10 (not connected, accepting connect from any host)
id=63 (not connected, accepting connect from any host)
id=127 (not connected, accepting connect from any host)
id=192 (not connected, accepting connect from any host)
id=228 (not connected, accepting connect from any host)
id=255 (not connected, accepting connect from any host)

Could you try the newest source code? For I can't reproduce the bug from newest source codes.

Thanks,

/Guangbao

No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".