Bug #11912 alter table leaves #sql-1efe_5 temp files around

• Submitted: 13 Jul 2005 13:11
• Modified: 5 Apr 2006 7:23
• Reporter: Ian Duggan
• Status: Won't fix
• Category: MySQL Cluster: Cluster (NDB) storage engine
• Severity: S2 (Serious)
• Version: 4.1.12
• OS: Linux (Linux 2.6.9 i686 (Debian))
• CPU Architecture: Any

[13 Jul 2005 13:11] Ian Duggan

Description:

alter table answers engine=ndb;

fails when the cluster does not have enough resources, and can sometimes leave temp #sql-1efe_5 tables around.

How to repeat:

2 ndbd nodes, 2GB memory each.

[NDBD DEFAULT]
NoOfReplicas=2                   # Number of replicas
DataDir=/var/lib/mysql-cluster   # Default DataDir
DataMemory=1024M                 # How much memory to allocate for data storage
IndexMemory=512M                 # How much memory to allocate for index storage
MaxNoOfConcurrentOperations=131052
RedoBuffer=96M
#TimeBetweenLocalCheckpoints=18

[NDB_MGMD]
Hostname=204.11.217.85          # Hostname or IP address of MGM node
DataDir=/var/lib/mysql-cluster  # Default DataDir

[NDBD]
hostname=204.11.217.80          # Hostname or IP address
[NDBD]
hostname=204.11.217.81          # Hostname or IP address
[MYSQLD]
hostname=204.11.217.71          # Hostname or IP address
[MYSQLD]
hostname=204.11.217.72          # Hostname or IP address
[MYSQLD]
hostname=204.11.217.73          # Hostname or IP address

alter table run on table of several million rows.

[13 Mar 2014 13:34] Omer Barnir

This bug is not scheduled to be fixed at this time.