Hey,

The perl script that does the backup issues a SLAVE STOP just before
it starts dumping. It also grabs SHOW SLAVE STATUS, which has a bunch
of file positions and I'm pretty sure it's everything that is in the
master.info file.

The backup I'd be pulling is going to be at least a day old, so it
will be out of sync and reseting the master will not help.

Thanks.

On Fri, 8 Oct 2004 13:24:00 -0400, [EMAIL PROTECTED]
<[EMAIL PROTECTED]> wrote:
>  
> Gary, 
>  
>     We go through the process of removing the slave DB and restoring from
> backup on a fairly regular basis (due to testing new functionality of our
> app). My first question would be about your backups - how are you doing
> them? 
>  
>      If you are doing a filesystem backup (taring the entire mysql data
> directory and replication files, for instance) then your master info and
> relay-log will have the information on where the to start the replication
> from the master. This is what we are doing. 
>  
>     I'm not 100% certain about using a mysqldump type program, but I suspect
> that you would need to reset the master logs after the backup to tell the
> slave to basically start from line 1. I dont know how you would ensure that
> the master would reset at the very last command that was backed up on the
> slave, perhaps someone using this type of slave/backup scenario could share
> some knowledge on the correct procedure. 
>  
>  
> Regards, 
>  
> Scott Tanner 
> Systems Administrator 
> Rowe/AMi 
>   
>  
>  
>  
>  "Gary Richardson" <[EMAIL PROTECTED]> 
> 
> 10/08/2004 01:01 PM 
> Please respond to Gary Richardson 
>          
>         To:        [EMAIL PROTECTED] 
>         cc:         
>         Subject:        Restarting Replication from Backup 
>  
>  
> 
> 
> Hey guys,
>  
>  I'm running a master/slave setup with v4.0.20. There are a hand full
>  of databases being replicated. For backups, we stop replication on the
>  slave and pull a text dump using mysqldump. I also record a 'SHOW
>  SLAVE STATUS' from the time of the backup.
>  
>  My replica server crashed last night. It looks like it had something
>  to do with the disk cache as the replica was trying to replay already
>  committed transactions (lots of duplicate record errors).
>  
>  After running an integritty check on the servers, the row counts are
>  out of sync for far more large tables than I care to manually fix. I'm
>  thinking of:
>  
>  1) deleting all the data on the replica
>  2) pulling a backup from a few days ago and re-importing it with
>  replication disabled on the replica (ie, comment out all replication
>  configuration directives).
>  3) artificially recreating master-info-file using the information from
>  'SHOW SLAVE STATUS'
>  4) restart the replica with replication turned back on
>  
>  With MySQL's two phase replication, will the IO thread automatically
>  figure out what file to start downloading and where to resume?
>  
>  Thanks.
>  
>  -- 
>  MySQL General Mailing List
>  For list archives: http://lists.mysql.com/mysql
>  To unsubscribe:   
> http://lists.mysql.com/[EMAIL PROTECTED]
>  
>  
>  
>  
>

-- 
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:    http://lists.mysql.com/[EMAIL PROTECTED]

Reply via email to