Replica stuck in DEAD / DYING / TRANSITION MOVE state

  • 7003110
  • 27-Apr-2009
  • 26-Apr-2012

Environment

Novell NetWare 5.0
Novell NetWare 4.11

Situation

Replica stuck in DEAD / DYING / TRANSITION MOVE state
NDS Manager reports the server in a TRANSITION MOVE State
DSREPAIR  on the server reports the replica in a DEAD / DYING State

Resolution

Possible Solutions.
1.  If the server affected has a Subordinate Reference that is stuck in a Dead state, make sure the server holding the dying S/R does not contain a copy of the parent.  If is does not, make sure the replica rings agree that this server is not in the replica ring.  This includes the server holding the dying S/R.  If it holds itself in the replia ring, then remove it from the ring in DSRepair.  Then run a Local Database Repair on the server holding the dying S/R replica.  Start DSTrace with +Part, +Sync, +J and let it run for a while.  This will most likely  remove the dying S/R replica.

If the above does not remove the dying SR then, after verifying synchronization status on ALL servers in the replica ring via DSRepair,  simply Add a Read-Write Replica of the partition to that server.   Once it goes to a RW ON state, you can Delete the RW replica and it will either remove the replica completely if it does not hold a copy of the parent partition or it will turn it back into a Subordinate Reference in and ON state.

2.  If the Server affected has a Read Write replica that is stuck in a Dead state, then preform the following

A.  On that server Load DSREPAIR -A | advanced options | Replica and partition operations | <Replica in the Dead state> | Destroy Selected Replica.   This will turn the replica into a S/R in a NEW state.  
B.  On the master replica of the partition, Load DSREPAIR -A | advanced options | Replica and partition operations | <Replica in the Dead state> | View replica ring | Select the server that is in a Dead state | Remove this server from the replica ring.
C.  Use NDS Manager to add a RW replica back on the server

3.  Try rebooting the server.  If the problem is communication, rebooting the server may get the communication functioning again (winsock problems, etc)
NOTE:  If it is the only replica left on the server, then using the DSREPAIR -XK2 procedure may be a possible option.  See TID #7001592 - Manually Removing All Replicas From a Server; DSREPAIR -XK2 for more information on the XK2 procedure.

Additional Information

Formerly known as TID# 10055496