Router crashes when I access Samba server over network

[WilhelmT]

I have a problem which seems to get worse in time. It could actually indicate more a hardware problem, but hopefully somebody has a simple solution.

My set up is the following:
I have an Orange Livebox, which is connected with ADSL.
Connected to this Livebox, with a cable, is a small fanless home server (Ubuntu Server) which runs MPD and a Samba server.
I also have two laptops (mac and win7) which connect wireless to the Livebox

I have no problems when I connect to the home server with Putty or use an MPD client. Everything is very stable and works very nice.
However, if I connect with my Macbook or Windows laptop to the Samba server and try to move/copy/view files the whole network goes almost always directly completely down and doesn't recover. The only thing that I can do is to unplug the power cable of the Livebox and put it back in to let it reboot.

Somebody told me it could have something to do with the dhcp settings of my server, but reading more about it doesn't give me really a clue. Should I really look more into this possibility or is it unlikely that this is the problem since MPD runs stable for a long time?
Anybody any clue what might be going on?

[Rubberman]

What are the IP addresses that the router and server are using?

[WilhelmT]

Router/Livebox: 192.168.1.1
Server 192.168.1.13

[Rubberman]

Is the server using DHCP to get its address from the router, or is it a static address? If static, what is the range of addresses assigned to dhcp in the router?

[WilhelmT]

I have this in my sudo vi /etc/network/interfaces

auto eth0
iface eth0 inet dhcp

[Rubberman]

What Linux distribution+version+kernel is your server running? Also, note that servers should properly use static addresses.

[WilhelmT]

I'm using Ubuntu Server 10.04.

So if I understand you correct I should change the /etc/network/interfaces in such a way that it uses a static IP address?

[Rubberman]

I'm using Ubuntu Server 10.04.

So if I understand you correct I should change the /etc/network/interfaces in such a way that it uses a static IP address?

That would be preferable; however, you will need to look at your router ( which I assume is your dhcp server also ) to see what range of addresses it has allocated to local static addresses, and if it hasn't, then you may need to adjust the range of dhcp addresses it allocates in order to allow for static addresses. For example, my router has been configured to allocate dhcp addresses in the range of 192.168.1.100 thru 192.168.1.253, allowing static addresses below 192.168.1.100. That way, I always know what is the actual IP address of all my servers, such as my network storage units, web servers, or any virtual servers that I need to access from a workstation or laptop.

[WilhelmT]

Okay. I changed the server to use a static IP address instead of using dhcp. It is working well, but I still don't seem to get rid of the crashing router problem.

I'm thinking about trying another router instead of the LiveBox, but I still feel that something could be wrong with the Samba settings, because the only way I get the LiveBox to crash is when I access (large) files via Samba. MPD clients on the network and login via putty work fine.
I looked a bit around in the logs. The first thing I noticed was that Samba couldn't access the CUPS. I fixed this by adapting the samba printer config. Error disappears, but the router is still crashing.

Other things that appear in the logs around and before the crash are those:
/var/log/samba/log.smbd:
[2011/10/01 12:35:21.677705, 0] lib/util_sock.c:474(read_fd_with_timeout)
[2011/10/01 12:35:21.677896, 0] lib/util_sock.c:1441(get_peer_addr_internal)
getpeername failed. Error was Transport endpoint is not connected
read_fd_with_timeout: client 0.0.0.0 read error = No route to host.
[2011/10/01 12:35:21.678226, 1] smbd/service.c:1251(close_cnum)
pc6 (192.168.1.11) closed connection to service share
[2011/10/01 12:35:21.719772, 1] smbd/server.c:267(remove_child_pid)
Scheduled cleanup of brl and lock database after unclean shutdown
[2011/10/01 12:35:41.739771, 1] smbd/server.c:240(cleanup_timeout_fn)
Cleaning up brl and lock database after unclean shutdown

Before the crash I also see this in syslog:
Oct 1 12:08:29 UNARMED winbindd[902]: [2011/10/01 12:08:29.150835, 0] winbindd/idmap.c:201(smb_register_idmap_alloc)
Oct 1 12:08:29 UNARMED winbindd[902]: idmap_alloc module tdb already registered!
Oct 1 12:08:29 UNARMED winbindd[902]: [2011/10/01 12:08:29.151043, 0] winbindd/idmap.c:149(smb_register_idmap)
Oct 1 12:08:29 UNARMED winbindd[902]: Idmap module passdb already registered!
Oct 1 12:08:29 UNARMED winbindd[902]: [2011/10/01 12:08:29.151200, 0] winbindd/idmap.c:149(smb_register_idmap)
Oct 1 12:08:29 UNARMED winbindd[902]: Idmap module nss already registered!
Oct 1 12:08:29 UNARMED winbindd[902]: [2011/10/01 12:08:29.151526, 0] winbindd/idmap_tdb.c:287(idmap_tdb_open_db)
Oct 1 12:08:29 UNARMED winbindd[902]: Upgrade of IDMAP_VERSION from -1 to 2 is not possible with incomplete configuration

Does this give any clues?

[Rubberman]

Personally, I have never heard of a protocol stream causing a router to crash. The only thing that comes to mind is that the router is for some reason filtering the packet data and has a bug in the filtering software. You might get a clue from the router configuration tools. As for the samba error log data, I'm not that deep of a samba expert to tell you what may be going on there.