[Samba] problems after lowering tombstone...

[mourik jan heupink - merit]

Hi all,

Seems we're in much more trouble now... Yesterday I lowered the dn 
stombstone lifetime from 180 to 60, to reduce the size of 
DC=DomainDnsZones, and ever since, things have become terrible...

The situation now:

I have separate fileserver, and three DC3's, all on the same subnet, kvm 
virtual machines, all running sernet-samba. (4.1.7 and 4.1.9)

== on DC3 ==
DNS queries for my samba.company.com still work

showrepl: ERROR(runtime): DsReplicaGetInfo of type 0 failed - 
(-1073610723, 'NT_STATUS_RPC_PROTOCOL_ERROR')

cpu load is not unusually high

== on DC2 ==
showrepl:
==== INBOUND NEIGHBORS ====

ERROR(runtime): DsReplicaGetInfo of type 0 failed - (-1073610723, 
'NT_STATUS_RPC_PROTOCOL_ERROR')

cpu load is high

DNS queries for my samba.company.com no longer work, but samba is 
listening on port 53, confirmed by netstat -tulpn

== on DC1 ==
Seems to be in even worse shape:
root at dc1:~# samba-tool drs showrepl
s4_tevent: EPOLL_CTL_DEL EBADF for fde[0x11a6020] mpx_fde[(nil)] fd[8] - 
disabling
ERROR(<class 'samba.drs_utils.drsException'>): DRS connection to 
dc1.samba.company.com failed - drsException: DRS connection to 
dc1.samba.company.com failed: (-1073741643, 'NT_STATUS_IO_TIMEOUT')

So... any tips how on earth to proceed now? Fortunately, the clients 
(mostly win7) can still connect, so the system is not completely down...

Any tips at all?