<p>The slave SQL thread failed because it could not create an internal conversion table while applying a replicated statement.</p>
<p>MySQL Error 1678: ER_SLAVE_CANT_CREATE_CONVERSION happens when the replication slave cannot create a temporary conversion table, usually due to missing disk space, wrong slave_load_tmpdir path, or file-system permissions. Free space or set a valid slave_load_tmpdir to resolve the issue quickly.</p>
Can't create conversion table for table '%s.%s'
The MySQL replication slave raises error 1678 when it tries to build a hidden conversion table to reconcile character sets or storage formats and fails.
The failure aborts the SQL thread, stops replication, and leaves the slave behind the primary, risking data divergence.
The error surfaces while the slave SQL thread replays statements that require character-set changes or ALTER TABLE operations.
It frequently appears after LOAD DATA or DDL statements originating on the primary with different character sets or collations than those on the replica.
Lack of writable space in the directory defined by slave_load_tmpdir or tmpdir prevents MySQL from creating the conversion table.
Incorrect permissions or a non-existent path for slave_load_tmpdir also trigger the error.
Filesystem limits such as exhausted inodes, disk quotas, or SELinux/AppArmor rules can block file creation.
Verify that the directory referenced by slave_load_tmpdir exists, has ample free space, and is writable by the MySQL user.
Update slave_load_tmpdir to a healthy location and restart the SQL thread.
If tmpdir is used instead, apply the same checks and increase the available storage or adjust permissions.
Cloud servers with small ephemeral disks often fill up during bulk loads; allocate a larger volume and move slave_load_tmpdir there.
After server migrations, old slave_load_tmpdir paths may vanish; recreate the directory and grant correct ownership.
Monitor disk utilization and inode counts on replicas continuously and alert when thresholds exceed 80 percent.
Configure slave_load_tmpdir on a dedicated partition with generous space and noexec safeguards removed.
Running replication health checks from Galaxy’s SQL editor lets engineers quickly spot a stopped slave thread and inspect global status variables without SSH access.
If the path set in slave_load_tmpdir does not exist or lacks write permissions, MySQL cannot create the conversion table.
A full volume or depleted inode count blocks file creation inside the temporary directory.
SELinux or AppArmor profiles may deny the mysqld process permission to create new files in the configured directory.
Replica volumes mounted read-only after a crash or maintenance make all write attempts fail.
Occurs when a table file format is incompatible between master and slave.
Indicates the slave could not read the master's binary log, often due to log purging.
Raised when the server cannot allocate enough memory or disk space for an operation.
Replication halts because the SQL thread cannot build a required conversion table. Without the table, the current event cannot be executed safely.
Ignoring the error risks data drift between master and slave. Always fix the root cause and restart replication.
No full server restart is needed. Setting the variable at GLOBAL scope and restarting the SQL thread is sufficient.
Galaxy lets you schedule SHOW SLAVE STATUS queries and surface stopped threads directly in the editor so you can act before lag grows.
.avif)
.avif){kind=logo}
