Archived
This forum has been archived. Please start a new discussion on GitHub.
Is the IceGrid registry a single point of failure?
in Help Center
One design consideration that we are looking at is trying to avoid any single point of failure.
Effectively we don't want a hardware failure on any single machine to take down the distributed application.
If for example the machine that we have our IceGrid registry on has a disk failure or other catastrophic harware problem and goes down - what are our options? If the IceGrid is being used as a locator, then one machine going down is going to mean that the clients are no longer able to find the servers, and the app grinds to a halt.
From what I have read there is only one IceGrid registry. Is there any way to have a backup? Of any kind?
Effectively we don't want a hardware failure on any single machine to take down the distributed application.
If for example the machine that we have our IceGrid registry on has a disk failure or other catastrophic harware problem and goes down - what are our options? If the IceGrid is being used as a locator, then one machine going down is going to mean that the clients are no longer able to find the servers, and the app grinds to a halt.
From what I have read there is only one IceGrid registry. Is there any way to have a backup? Of any kind?
0
Comments
-
Hello Tim,
The IceGrid registry is indeed a single point of failure right now and there's no easy way to setup a backup and have the clients automatically use this backup if the main registry isn't available. We could implement a replicated version of the registry however. If you have a commercial need for this, please contact us at info@zeroc.com.
Cheers,
Benoit.0