NSM 3.0 - latest version - running on OES2 SP3 64 bit
Running one engine, one agent and one server, all on the same server which has a read-write replica.

When deleting edir groups that have collaborative storage policies I am getting two issues. The first is that eDirectory on the server containing the storage is crashing. The message in /var/log/messages is

Jan 13 11:09:52 files kernel: ndsd[10781]: segfault at 0000000000000121 rip 0000
2aaaaab03fde rsp 0000000041806f60 error 4

I can do an rcndsd start and it will load again, and then the folder deletion continues.

This was also happening with version 2.5, and I was hoping the upgrade would fix it, but although it seems to happen less often (ie only once while removing about 7 groups with around 20 members each, each member having a personal folder under the group folder, whereas previously it would happen several times) it is still happening.

It may have started after applying OES 2 SP3, but I can't be sure as we haven't deleted groups since about July last year.

The eDir crash only occurs with NSM created folders - an equivalent number of "ordinary" folders created with a script can be deleted without issue.

I have changing the server the collaborative storage is created on but get the same issue.

The other issue is after the deletion I have two transactions that don't complete. The action is "deprovision collaborative storage" and the action state is "waiting for ds sync". The current error is "17 : Object attribute has too many values.". I've checked DS sync and all servers are synchronised. The server with the storage on it has a read-write replica. But I got the same results (ie both problems) when it didn't have a replica. I put one on today to see if it fixed the problems.

Any advice appreciated.