cancel
Showing results for 
Search instead for 
Did you mean: 

24x7 activiti database?

mingway
Champ in-the-making
Champ in-the-making
Hi,

We need high availability activiti engine, 24x7. If we bring down activiti database(oracle) for maintenance, and switch to a backup db, how can we handle those in-progress process instances in the down database? How can we move the running data to the backup db?

Is there any mechanism?  I think basically we need 1 activiti engine using 2 databases. How to share running data in the 2 dbs? Or can activiti engine use database in the cloud?

Thank you very much,
5 REPLIES 5

jbarrez
Star Contributor
Star Contributor
I think the easiest is to do it on the DB level. I only have experience with mysql on that front, but there we used a setup with master-slave replication where we could divert traffic to one of the slaves when we needed to take down the master. I don't know about Oracle, but I'm sure they'll have something similar.

mingway
Champ in-the-making
Champ in-the-making
Thanks, jbarrez!

We do use shareplex to sync up data between 2 dbs. But even that small latency is  not acceptable. So still trying to find some solution.

jbarrez
Star Contributor
Star Contributor
I remember that I worked with a sysadmin before that used dns routing for that.
He switched all traffic to a certain db for a while, upgraded the other db (or whatever) and brought it back online and restored the dns.

But I'm no sysadmin. But I really think that's the area you should look into rather than on the application level.

laures
Champ in-the-making
Champ in-the-making
i'm currently evaluating activiti and have a similar problem:

we require application based replication capabilities. currently all data is replicated using jpa and eclipseLink.

is it possible to configure the persistence of activitiy to replicate its data to multiple connections?
Is it possible to implement a custom persistence mechanism?

Greetings,
Alexander

jbarrez
Star Contributor
Star Contributor
> is it possible to configure the persistence of activitiy to replicate its data to multiple connections?

No, not out of the box that is. I don't know if it exists for other databases, but I remember using in the past a driver for my database which did something similar for me.

> Is it possible to implement a custom persistence mechanism?

It is possible, it is open code after all … but I won't go down that path. It will take you a lot of time and will make it difficult to upgrade to new versions.