[Samba] core dump after live migrating virtualized environment

[Jeff Blasius]

Dear Samba-
We're using the virtualization application openVZ to run small web
services for researchers. Unlike other virtualization software, openVZ
uses the host file system for the guest OS; essentially the root file
system of the guest is a chrooted subdirectory on the host.

When openVZ migrates a guest between two host machines it uses rsync
to transfer the files. Unfortunately, this causes samba to panic and
not accept any new connections (established connections are OK). In
the logs I get "
tdb(/var/cache/samba/ntforms.tdb): tdb_reopen: file dev/inode has
changed!" . Of course the inode changed, but does it need to panic?
Is there a workaround?

Centos 5
samba-3.0.25b-1.el5_1.4


[2008/01/22 11:53:46, 0] lib/util_tdb.c:tdb_log(662)
 tdb(/var/cache/samba/ntforms.tdb): tdb_reopen: file dev/inode has changed!
[2008/01/22 11:53:46, 0] smbd/server.c:open_sockets_smbd(572)
 tdb_reopen_all failed.
[2008/01/22 11:53:46, 0] lib/util.c:smb_panic(1654)
 PANIC (pid 22099): tdb_reopen_all failed.
[2008/01/22 11:53:46, 0] lib/util.c:log_stack_trace(1758)
 BACKTRACE: 6 stack frames:
  #0 smbd(log_stack_trace+0x1c) [0x555555776ffc]
  #1 smbd(smb_panic+0x43) [0x5555557770e3]
  #2 smbd [0x55555582ac3a]
  #3 smbd(main+0x710) [0x55555582b370]
  #4 /lib64/libc.so.6(__libc_start_main+0xf4) [0x2aaaad3818a4]
  #5 smbd [0x5555555bbfe9]
[2008/01/22 11:53:46, 0] lib/fault.c:dump_core(181)
 dumping core in /var/log/samba/cores/smbd



-- 
Jeff Blasius / jeff.blasius at yale.edu
Phone: (203)432-9940  51 Prospect Rm. 011
High Performance Computing (HPC)
UNIX Systems Administrator, Linux Systems Design & Support (LSDS)
Yale University Information Technology Services (ITS)