cancel
Showing results for 
Search instead for 
Did you mean: 

Node reference collision on two independent instances ?

faizaan
Champ in-the-making
Champ in-the-making
Hi,

Are there any chances of having two different contents on two different independent alfresco instances have same node reference.

If yes, then is such case possible very frequently.


Thanks,
Faizaan
6 REPLIES 6

mrogers
Star Contributor
Star Contributor
Yes its possible.  NodeRefs are only valid in a particular repo.

However NodeRefs are fairly long guids so its not very likely to have duplicate node refs unless nodes have been transferred between repos.

What uniquely identifies a node is its repositoryId/storeId/docId

faizaan
Champ in-the-making
Champ in-the-making
Thanks mrogers for quick reply !

We have a use case where we have a production repository having around 100K contents, and we having another independent repository (migration instance )where we having migrated another around 100K content from legacy system.

Now we want to replicate content from this migration instance to production repository using replication services, so can we safely replicate that without affecting existing production data.

Node References of contents on production and migration will not be same ?

Thanks !
Faizaan

mrogers
Star Contributor
Star Contributor
Depending upon how the nodes are replicated then yes node refs can be the same.   For example the transfer api will preserve node refs.

faizaan
Champ in-the-making
Champ in-the-making
Yes, noderefs of contents will be preserved by transfer api for contents that are getting replicated and it is required.

But what question I have is while replicating content in target repository is there any chance that node reference of content that is getting replicated from source server will match to some other content on the target server that already existed (not replicated, but created by some user) on target server.

Thanks !
Faizaan

mrogers
Star Contributor
Star Contributor
They are uuids so generally there's time and machine id seeding the algorithm which makes overlap less likely.

Its possible to specify a node ref and there are a few nodes refs that are hard coded, so in those cases overlap is more likely.

faizaan
Champ in-the-making
Champ in-the-making
Thanks mrogers for your valuable inputs !

So I conclude that node refs will not overlap while we are replicating content from source to target repository excluding node refs that are hard coded.

Provided that both instances are running on different machines.

Thanks !
Faizaan