MySQL Bugs: #41292: GCP stop errors exits both data nodes

Bug #41292	GCP stop errors exits both data nodes
Submitted:	8 Dec 2008 10:18	Modified:	25 Mar 2009 11:06
Reporter:	Geert Vanderkelen	Email Updates:
Status:	Duplicate	Impact on me:	None
Category:	MySQL Cluster: Cluster (NDB) storage engine	Severity:	S2 (Serious)
Version:	mysql-5.1.30-ndb-6.3.20	OS:	Any
Assigned to:	Assigned Account	CPU Architecture:	Any

Description:
While testing bug #28622 I ran once in some temporary errors with the infamous GCP stop.. Why I didn't open new bug for that, no idea :(
Here is how to reproduce it.

How to repeat:
Basically, test case of #28622:

Three tables, all having the same DDL:

CREATE TABLE `tab1` (
  `id` char(12) NOT NULL,
  `timestamp` timestamp NOT NULL default CURRENT_TIMESTAMP on update CURRENT_TIMESTAMP,
  PRIMARY KEY  (`id`)
) ENGINE=ndbcluster DEFAULT CHARSET=ascii;

CREATE TABLE `tab1_newdata` (
  `id` char(12) NOT NULL,
  `timestamp` timestamp NOT NULL default CURRENT_TIMESTAMP on update CURRENT_TIMESTAMP,
  PRIMARY KEY  (`id`)
) ENGINE=ndbcluster DEFAULT CHARSET=ascii;

CREATE TABLE `tab2` (
  `id` char(12) NOT NULL,
  `timestamp` timestamp NOT NULL default CURRENT_TIMESTAMP on update CURRENT_TIMESTAMP,
  PRIMARY KEY  (`id`)
) ENGINE=ndbcluster DEFAULT CHARSET=ascii;

The small C application populating tables on bug #28622.
Then do the following concurrently:

shell_1> while true; do /data1/mysql/ndb-6.3.18/bin/mysql test -e "REPLACE INTO tab1 SELECT * FROM tab1_newdata"; done

shell_2> while true; do /data1/mysql/ndb-6.3.18/bin/mysql test -e "DELETE from tab2 where timestamp<(CURDATE()-1)"; done

Eventually both nodes will get temporary error.

Verified using MySQL Cluster 6.3.18 and 6.3bzr (latest pull)

GCP Stop

Attachment: ndb_error_report_20081208110944.tar.bz2 (application/octet-stream, text), 86.64 KiB.

Might be a simple load problem. Error message GCP stop is a bit generic in this repect.

http://bugs.mysql.com/bug.php?id=37227

I encounter this error on 5.1.44-ndb-7.1.3-cluster-gpl-log even with MaxNoOfExecutionThreads=8. What can I do to avoid this problem in the future ?

cluster log

Attachment: ndb_1_cluster.log (application/octet-stream, text), 0 bytes.

cluster error log

Attachment: ndb_5_error.log (application/octet-stream, text), 508 bytes.

cluster log file

Attachment: ndb_5_out.log (application/octet-stream, text), 51.66 KiB.

cluster trace log

Attachment: ndb_5_trace.log.1 (application/octet-stream, text), 274.79 KiB.