I've got a master database that is doing lots of inserts, deletes, and updates. It has a number of slaves hanging off of it. One of those slaves is completely unloaded - so it keeps very close track of the master. Let's call that machine slave1. My problem is doing restores from backups. The only way I've been able to get a reliable restore to occur is to do a FLUSH TABLES WITH READ LOCK on the master, then copy the /var/lib/mysql/* directories, do a SHOW MASTER STATUS on the master, and then UNLOCK TABLES on the master. Then after copying the snapshot to a new slave I do a CHANGE MASTER command on the new slave and everything works.
So far, so good.
The problem is that these databases are pretty large, and are regularly updated. Stopping the master to do a snapshot is a pretty drastic move. What I'd like to do is find a good recipe to use slave1 as the machine to snapshot. I've tried the following, but things don't ever seem to work:
On slave1:
mysql> SLAVE STOP; mysql> FLUSH TABLES WITH READ LOCK; Then I switch over to /var/lib/mysql/ and tarball the directory cd /var/lib/mysql tar zcvpPf /var/tmp/snapshot.tar.gz .
Then I copy over the snapshot over to the new slave, and do the following: /etc/init.d/mysql stop cd /var/lib/mysql tar zxvpPf /var/tmp/snapshot.tar.gz
Then, when the tarball is finished unpacking, I run:
/etc/init.d/mysql start
I then log in as mysql superuser, and run the command: mysql> slave start;
And what I get back in return is:
ERROR 1201: Could not initialize master info structure, check permisions on master.info
Everything is owned by mysql.mysql, the permissions on master.info are:
-rw-rw---- 1 mysql mysql 61 Oct 5 21:05 master.info
So what's going wrong?!? Help!
All of the machines are running x86/Debian Linux/MySQL 4.0.13.
Dave
-- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
