Not the answer you need?
Register and ask your own question!

pt-table-sync with no unique index - fails with no more space left

TeteluTetelu EntrantCurrent User Role Beginner

I have a master-slave replication and I'm having trouble syncing one table that has no unique index, so I used the --no-check-slave option and when I thought it all goes well, error occurred:
Error writing file '/dev/shm/MYGFvDuG' (Errcode: 28) [for Statement "SELECT `[B]column1[/B]`, `[B]column2[/B]`, `[B]column3[/B]`, COUNT(*) AS __maatkit_count FROM `[B]database[/B]`.`[B]table[/B]` WHERE 1=1 GROUP BY `[B]column1[/B]`, `[B]column2[/B]`, `[B]column3[/B]` ORDER BY `[B]column1[/B]`, `[B]column2[/B]`, `[B]column3[/B]` LOCK IN SHARE MODE"] at line 5301 while doing database.table on slave.server

The command I used is:
pt-table-sync --execute --no-check-slave h=localhost,D=database,t=table h=slave.server,u=user,p=password,D=database,t=table

The dry-run looks ok, no errors, version: pt-table-sync 2.1.1

I know the problem is that the query quickly eats all the ram, then goes writing in /dev/shm/ until it fills it, then fails with no more space left..
Mysql error log shows: [ERROR] /usr/sbin/mysqld: Sort aborted

Any solution to write it somewhere else? Or any other solution?

Any help will be really appreciated.
Sign In or Register to comment.

MySQL, InnoDB, MariaDB and MongoDB are trademarks of their respective owners.
Copyright ©2005 - 2020 Percona LLC. All rights reserved.