Re: OT: database clusering & failover
Re: OT: database clusering & failover
- Subject: Re: OT: database clusering & failover
- From: James Cicenia <email@hidden>
- Date: Wed, 1 Nov 2006 12:51:27 -0600
Hello -
I have a situation where I would like to publish subsets of my
database around the web.
Basically we have a master WebObjects database that does a lot. Then
we need to publish
data to remote databases. I was wondering if anyone knows a way to
just do a subset
of a table to each remote database?
Thanks
James
On Nov 1, 2006, at 8:23 AM, Kieran Kelleher wrote:
Hi Joshua,
I use MySQL and I like it!
I am just today revisiting/fine-tuning my replication/backup
strategy ("how to" notes soon to be posted on my blog http://
homepage.mac.com/kelleherk/iblog/) since I upgraded master mysql
server from mysql 4.0 on a G4 XServe to 4.1 on a dual G5 XServe
just this morning.
We have a half-dozen databases - generally using innodb transaction
engine by default for all project databases.
Hardware Failure Protected
---------------------------------
http://homepage.mac.com/kelleherk/iblog/C711669388/E351220100/
index.html
I use mysql replication ..... very easy to set up..... allows high-
availability on the master database. Instructions for 4.0 (still
valid) are on my blog. Improved, even easier replication setup for
mysql 4.1 blog entry is forthcoming.
Accidental data loss protected
------------------------------------
Currently: I have a script that shutdowns the replication slave
each night and makes a copy of the mysql data folder and then
restarts the replication slave. That data gets copied to firewire
backup drives that are rotated offsite weekly.
Planned: mysqldump rolling backups of individual databases from
replication slave keeping up to about 10 days of rolling backups.
High update/insert dbs will get dumped more frequently. Plan to use
rsync to automate offsite/remote storage of backups.
The nice thing about replication is that you can start/stop/lock
the slave for backups and your master db server just keeps on
running 100% 24/7.
We also have a little java program that updates a table with a
timestamp on the master and reads the same timestamp on the slave
and compares the two to see how far behind the slave is. If it
falls too far behind we get warning emails. So, if replication
fails and stops working (which it can occasionally), then the
emails start coming.
HTH, Kieran
On Nov 1, 2006, at 6:56 AM, Kuon - Nicolas Goy (Goyman.com SA) -
時期精霊 wrote:
On 1 Nov 2006, at 07:15, Joshua Paul wrote:
What solutions have people here used to deal with database or
hardware failures? Has anyone tried or deployed using MySQL
Cluster? If so, what pitfalls/hurdles did you encounter?
I never used mysql for serious project.
I have experience with postgresql and slony (with script to make
the failover database read write when the first database fail) and
oracle cluster.
Never had any serious problems.
Regards
--
Kuon
CEO - Goyman.com SA
http://www.goyman.com/
"Computers should not stop working when the users' brain do."
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
40mac.com
This email sent to email@hidden
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
40jimijon.com
This email sent to email@hidden
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden