nds out of sync after restore of server

What happened:
1. Make vmware snapshot of server.
2. Make OES-Upgrade 2018SP1 to SP2. Work without problems.
3. Remove snapshot and make new one.
4. Make Offline OES-Upgrade to 2023. Error at configuring ndsd (edir_oes2018_sp2 was with SERVICE_CONFIGURED="no")
5. Server Object was deleted from secondary server (Bad failure, I know...)
6. After several tries to configure nds, I aborted and reverted to snapshot.

Situation now:
nds on Master-Server runs and shows both servers in replica ring
Both servers are in timesync and are running ok but without nds-sync

On Secondary server replica ring shows only secondary server and no master server. ndsrepair -N shows only secondary

How to get master server in serverlist on secondary?
Is it possible to "copy" nds from master to secondary?

Please show me hints to solve this problem.

  • Does ndsrepair -S and Reset local schema so the Job I needed?

  • I know such situations from the field. In this case, an NDS crash can be assumed. As long as the master is still OK, there are various options. First, check with NDS Trace and corresponding commands whether the background sync of the NDS is still working at all, whether there are obituraries stuck somewhere that are preventing the sync and whether there is perhaps a timestamp (future) or transitive vector problem or the NDS sync fails with -672, -663, -625 or whatever. If something is already recognizable as a problem here, there is usually only one call to open. Depending on the problem, the issue can really be resolved, but not via a forum at the green table. As I said, many things can be analyzed with an NDS trace. It may be that an xk3 ndsrepair on the replica server can sweep out the external references and then be recreated via an ndstrace = *b. I know too little about the disaster and as I said, green table.

    Greetings George

  • No, this won't help at all.

    For clarification i'll try to sum up what has happened (and it's of vital interest that you carefully comment on each and every of the following lines:
    - it's a two server tree with just one partition [root]
    - the upgrades (from sp1 to sp2 and the attempt with 2023) were done on the master of [root]
    - after the failed attempt you deleted the NCP server object of the master of [root] from the tree
    - you then reverted the master of [root] to the sp2 snapshot
    - if you now run "ndsrepair -P" on the master and select "1" followed by "10" it does show you both servers
    - if you do the same on the "untouched" server you see just the "untouched" box with a r/w replica, the master does NOT appear
    ---
    Do you have some sort of backup of the "untouched" server with the r/w replica from a time where the master server's NCP server object hadn't been deleted yet? It could be a dsbk-based backup, a snapshot or a snapshot-based backup, even a file based backup would help. And it doesn't have to be a recent one. If you have one: just tell us here, but DON'T bring it online under NO CIRCUMSTANCES. Just tell us that you have one and we'll guide you through the process.
    If you have no backup but support, you can call them, of course, as it's not too hard to resolve the issue with dump.