SDS Faults / Sync clarification

[rikko]

Good evening,

I'm looking for clarification please. If I'm trying to manually sync between two admin nodes in a cluster, which way will the synchronization go?

Here's the background, system I'm managing comprises of 4 nodes. I have a bunch of numbers hosted on node #4, with node #3 as a secondary host. When I try to modify users associated to any of these numbers (doing so from a primary host), replication is ok between nodes #2, #3 and #4, but it fails on node #1. As a result, I get multiple users under the same number, on node #1, and as much SDS faults in the process.
I've done the form comparison, and of course a number of users were found to exist only on node #1 (850+ that shouldn't be there). Naturally, I want to initiate sync manually, but it's unclear to me which way the synchronization will go. I don't want to copy these node #1 nonexistent users to other nodes, that's my concern. Ideally, I want the system to delete the extra users on node #1, automagically, instead of me going and removing them manually (which also works).

Can anyone clarify this for me please?

Thanks for reading

[ralph]

It will push to whichever system host a particular extension.
So if the extension is hosted on system 4 then it will push the changes to system 1.

You can try this:
On one of the extensions that has an issue, delete it then recreate it. Sometimes that will fix it.
But I wouldn't be afraid of doing the sync although that may not fix it either. You'll have to figure out what's causing it to fail on node 1.

Ralph

[rikko]

Hi Ralph,

thanks for the clarification. When you say to try to delete one extension, I presume you mean from the primary host, yes?

Re. the failure, I'm not sure what had caused it to fail initially. Now I get SDS errors piling up with every user change on node #3. These faults are that "...there can be only one prime user..." or that "...user does not exist locally. Synchronize data first...". To give you more background, we update extension user names to match the name of the occupant. This name update is pushed from an external app, developed by some smart ex-mitelians. So the change is pushed to nodes #3 and #4, and SDS sharing takes care of nodes #1 and #2. I suspect that when distribution initially failed, it wasn't taken care of immediately, and what I'm facing now is consequence of that. I'll try what you mentioned, let's see how successful will I be.

Thanks

[rikko]

Sync worked! Users fully match now.

I'm chuffed.

Thanks Ralph

[rikko]

And further confirmation, any consequent edits of users on node #4 did not result in new SDS faults.

For interested parties, what I've done is:
- logged into the primary host for these users in question
- opened Network Elements page
- clicked on Sync between the nodes in question
- chosen the Form that needed syncing (Users in this case)
- chosen Overwrite for sync properties
- started syncing

Hope this helps