How reliable are modern databases in the presence of disk errors?

views:

answers:

How reliable are modern databases in the presence of disk errors?

InterBase had an architecture that caused disk-writes to leave the database in an always-consistent state -- 97 things every software architect should know, p87

Is this property finally common in 2010 ?

Is there a study about database stability/reliability in presence of disk errors ?

+2 A:

The most reliable database is, without a doubt DB2. But not that toy LUW thing you run under UNIX and Windows.

I'm talking about the big grunter, DB2/z, the one that runs on the mainframe.

Quake in terror, all you puny, non-mainframe, database wannabes :-)

By the way, the reason it's so reliable is the underlying hardware. Non-mainframe platforms can only dream of the sorts of redundancy levels and self-checking that goes on in the System z boxes.

paxdiablo 2010-09-08 08:24:33

a commodity hardware cluster has high redudancy. for lower price ?

2010-09-08 09:03:16

@user: a well designed, assembled, configured and maintained cluster has a enough redundancy to be compared to big iron. A lot of times the comparison goes "While cheaper than a mainframe a cluster still doesn't have...", but it is close enough that comparing them makes sense. The thing is the guys who can set it up that way are as rare and expensive as big iron sysadmins, and it is easy to mess up.

dmckee 2010-09-08 18:45:39

Interbase is still alive and Interbase free fork is Firebird

Hugues Van Landeghem 2010-09-08 17:15:13

ansaurus

tags:

views:

answers:

How reliable are modern databases in the presence of disk errors?

related questions