On 8/22/06 4:50 AM, "Tony Baker" <email@hidden> wrote:
I have an Xserve running 10.4.6
It is running windows services and is set up as a "windows
standalone
server"
The windows users are mapping the one directory shared by the
server.
The windows services stop everynow and again with
Got SigTerm: Going Down in /var/log/samba/log.nmbd
In one situation where I was seeing this issue it was due to
conflicting
Master Browsers on the network. By moving to a WINS environment
and
disabling File Sharing on some of the workstations we were able
to stop the
Master Browser election process from causing the SigTerm: Going
Down events.
From reading some of the other discussions it seems unlikely
that this is
the cause for everyone's issues but it might help some users.
Similar to
using WINS you could promote the server to a PDC and log the
Windows
stations into its domain, if you are finding that Master Browser
elections
are causing your issue. The logs of Windows stations could help
determine
if that's the case. Be wary of this as it might not be your
issue, so
verify it is before going through all the work of moving to a
WINS or PDC
environment.
This particular machine on which this problem has occurred is set
up as a "standalone server", I cannot make this machine a PDC on
this network,
but I have other machines varying from 10.4.3 upto 10.4.6 and I
see this problem on machines configured as a PDC as well.
Although this doesn't happen frequently, it is happening enough
to be
a complete pain now.
I can see others are seeing the same problems. (Apple Discussions)
Does anyone have any workarounds or better still cures?
What about updating to 10.4.7? Is anyone seeing these issues when
running 10.4.7
I have seen continued reports of this in 10.4.7. I have not
personally
experienced this in 10.4.7.
Has anyone else any experience of Windows service stopping with
"Got Sigterm: Going Down" on 10.4.7
One good piece of advice I can give, is to take anything from the
Apple Discussions with a few bags of salt.
What shows up in your system.log for that same time-frame ?
If it is a master browser contention issue, you can edit /etc/
smb.conf on the Macs and add:
local master = no
in the [global] section
Although, why scuttle a client-server setup by running filesharing
on any/various/numerous client machines ?
On Windows clients, edit the registry, search for
"MaintainServerList" - change it from "Auto" to "No"
Using the Administrative Tools, disable "Computer Browser"
Thanks for taking the trouble to reply
Below is part of the system.logs approx around time the problem
occurred
==========================================
Aug 22 11:02:24 sybserver kernel[0]: netsmb_dev: loaded
Aug 22 11:16:28 sybserver crashdump[4333]: servermgrd crashed
Aug 22 11:16:28 sybserver launchd: com.apple.servermgrd: exited
abnormally: Segmentation fault
Aug 22 11:16:28 sybserver launchd: com.apple.servermgrd: 9 more
failures without living at least 60 seconds will cause job removal
Aug 22 11:16:29 sybserver servermgrd: servermgr_dns: hostname and
DNS entries for this server are synchronized
Aug 22 11:16:29 sybserver crashdump[4333]: crash report written
to: /Library/Logs/CrashReporter/servermgrd.crash.log
Aug 22 13:38:27 sybserver crashdump[6464]: servermgrd crashed
Aug 22 13:38:27 sybserver launchd: com.apple.servermgrd: exited
abnormally: Segmentation fault
Aug 22 13:38:27 sybserver launchd: com.apple.servermgrd: 9 more
failures without living at least 60 seconds will cause job removal
Aug 22 13:38:27 sybserver servermgrd: servermgr_dns: hostname and
DNS entries for this server are synchronized
Aug 22 13:38:27 sybserver crashdump[6464]: crash report written
to: /Library/Logs/CrashReporter/servermgrd.crash.log
Aug 22 14:38:58 sybserver crashdump[7376]: servermgrd crashed
Aug 22 14:38:58 sybserver launchd: com.apple.servermgrd: exited
abnormally: Bus error
Aug 22 14:38:58 sybserver launchd: com.apple.servermgrd: 9 more
failures without living at least 60 seconds will cause job removal
Aug 22 14:38:58 sybserver /Applications/Server/Server Admin.app/
Contents/MacOS/Server Admin: [4165] ServerManager session failed in
connect(sybserver.company.com,192.10.200.26,311): 61\n
<snip>
Aug 22 17:39:27 sybserver kernel[0]: nmbd[9996] uses send/recv on a
pipe
Aug 22 17:43:09 sybserver crashdump[10070]: servermgrd crashed
Aug 22 17:43:09 sybserver launchd: com.apple.servermgrd: exited
abnormally: Segmentation fault
Aug 22 17:43:09 sybserver launchd: com.apple.servermgrd: 9 more
failures without living at least 60 seconds will cause job removal
Aug 22 17:43:10 sybserver servermgrd: servermgr_dns: hostname and
DNS entries for this server are synchronized
Aug 22 17:43:10 sybserver crashdump[10070]: crash report written
to: /Library/Logs/CrashReporter/servermgrd.crash.log
[2006/08/22 14:38:58, 0] /SourceCache/samba/samba-92.19/samba/
source/smbd/service.c:make_connection_snum(620)
'/tmp' does not exist or is not a directory, when connecting to
[IPC$]
[2006/08/22 14:38:58, 0] /SourceCache/samba/samba-92.19/samba/
source/lib/util_sock.c:get_peer_addr(1016)
getpeername failed. Error was Socket is not connected
[2006/08/22 14:38:58, 0] /SourceCache/samba/samba-92.19/samba/
source/lib/util_sock.c:get_peer_addr(1016)
getpeername failed. Error was Socket is not connected
[2006/08/22 14:38:58, 0] /SourceCache/samba/samba-92.19/samba/
source/lib/util_sock.c:write_socket_data(446)
write_socket_data: write failure. Error = Broken pipe
[2006/08/22 14:38:58, 0] /SourceCache/samba/samba-92.19/samba/
source/lib/util_sock.c:write_socket(471)
write_socket: Error writing 4 bytes to socket 23: ERRNO = Broken
pipe
[2006/08/22 14:38:58, 0] /SourceCache/samba/samba-92.19/samba/
source/lib/util_sock.c:send_smb(663)
Error writing 4 bytes to client. -1. (Broken pipe)
Any assistance would be appreciated
I cannot do anything to the windows clients as they are out my control
If you post this to the list, many other very knowledgeable people
can also offer help.
From those entries, next stop will be to look at /Library/Logs/
CrashReporter/servermgrd.crash.log
However, the "segmentation fault" does not look encouraging, and
could mean a hardware problem.
Boot the server from the install CD and use Disk Utility to verify/
repair your system drive (assuming you do have a known-good backup).
Then download memtestosx , http://www.memtestosx.org/
and test your RAM overnight, when the server is not in use (shut off
services).
You may wish to run the Apple diagnostics that should be on one of
the disks that came with your server.
By the way, is this 10.4 Server and an Xserve ? Some hardware and Mac
OS specs/info would be good to share as well.
