Just to update you all here is what I've decided to do. Since the boxes are on the same subnet. Since the boxes will replace each other. Since there is so much data ( 1.2TB ).
Use rsync w/o ssh. Rsync the smallest idb files. Stop the DB and rsync -W the largest files which account for about 800+GB. Takes about 4+ hours. Then rsync the smaller files again. Takes about an hour to get the deltas. -W says don't do deltas just copy the file new. No drdb but yes LVM. On Wed, Sep 7, 2011 at 4:25 PM, Anthony Acquanita <[email protected]> wrote: > If I owned the equipment I would put the 4 drive raid in a bigger box > with more drives. But I don't trust our colo to do the work. They > also setup the mysql partition w/o LVM, idiots. > > A few people have assured me that rsync'ing multiple times during and > after the db is running and then stopped will work. > > I have about 1.2TB of data to move. I have yet to be convinced that a > mysql dump and restore is worth the time. If the rsync fails it's my > last option. > > In the future a enterprise san would be great or a clustered fs. > > On Wed, Sep 7, 2011 at 2:02 PM, Matthias Johnson > <[email protected]> wrote: >> I haven't really had to do this yet with databases but this obviously is >> done all the time and I have had an interest in learning it but haven't >> taken the time to test this kind of scenario. This can't be accomplished >> with MySQL Replication? >> http://dev.mysql.com/doc/refman/5.0/en/replication.html or Synchronous >> Replication? http://dev.mysql.com/doc/refman/5.0/en/mysql-cluster.html >> >> -- >> Matthias A. Johnson >> matthias.a.johnson aut gmail dot com >> >> >> On Tue, Sep 6, 2011 at 8:23 PM, Anthony Acquanita <[email protected]> wrote: >>> >>> Hello everybody! >>> >>> So I'm moving a rather large mysql DB from one server to another. It >>> sucks. I can't have any real long downtime so dump/restore, not an >>> option. >>> >>> It's innodb so I can't rsync and repair. What I'm using now is >>> percona xtrabackup which gives me a window of about three hours from >>> read lock to recovery. Not too bad for a 800+GB DB. >>> >>> Never mind all that. The real question is, "What can I do now to >>> avoid this in the future?" I don't have a second box yet so a slave >>> won't help. This box is already a hybrid master/slave setup as it is. >>> This leave me with DRDB. >>> >>> Has anyone used DRDB with LVM and Mysql? Thoughts? Concerns? >>> What I would really like to work is drdb on the box that exists now >>> and then when I get a new box in the future setup drdb on that and >>> slave to it. >>> >>> Thanks. >> >> _______________________________________________ >> Mid-Hudson Valley Linux Users Group http://mhvlug.org >> http://mhvlug.org/cgi-bin/mailman/listinfo/mhvlug >> >> Upcoming Meetings (6pm - 8pm) MHVLS Auditorium >> Sep 7 - DIY 3D Printing and the Makerbot Thing-o-Matic >> Oct 5 - Distributed Authentication Systems >> Nov 2 - Nov 2011 >> >> > _______________________________________________ Mid-Hudson Valley Linux Users Group http://mhvlug.org http://mhvlug.org/cgi-bin/mailman/listinfo/mhvlug Upcoming Meetings (6pm - 8pm) MHVLS Auditorium Oct 5 - Distributed Authentication Systems Nov 2 - Nov 2011 Dec 7 - Chef
