Karandeep Singh
2006-Mar-29 10:59 UTC
[Samba] Linux Samba server mounts hundreds of filesystems
We are a fairly large site with several thousand unix filesystems available for a Samba server to mount. Some Windows user or users are running an application or command that causes the Samba server to mount everything. Even though we have a specially hacked kernel that allows upward of 7000 mounts, the mount table fills up and messes up the automounter. This is not a Samba issue. I would like some pointers as to how to identify the command(s) users are running on the Windows side that causes this on our Linux Samba server. We have hundreds of Samba clients and increasing the debug level for all does not seem the way to go. I am also not sure as to which the suspect clients are to zero in on them. Any thoughts or suggestions as to how to proceed with the debug, what log level to set, and what strings to look for in (which) log files would be appreciated. Here are some details. uname -a Linux plxs0131 2.4.9-45lxset36tcpsmp #1 SMP Wed Jan 12 09:16:34 PST 2005 i686 unknown smbd -V Version 3.0.10 Thanks, -KD
Ben Walton
2006-Mar-29 13:58 UTC
[Samba] Linux Samba server mounts hundreds of filesystems
You could use the per machine configuration options to enable a different debug level for only a few clients. Something like (in [global]): include = /etc/samba/smb.conf.%m Then create a file called /etc/samba/smb.conf.<client machine> with the debug settings there. I'm not sure if doing this requires a config file for every machine or not though. Just a thought. -Ben On Mon, 2006-03-27 at 16:18 -0800, Karandeep Singh wrote:> We are a fairly large site with several thousand unix filesystems available > for a Samba server to mount. Some Windows user or users are running an > application or command that causes the Samba server to mount everything. Even > though we have a specially hacked kernel that allows upward of 7000 mounts, the > mount table fills up and messes up the automounter. This is not a Samba issue. > I would like some pointers as to how to identify the command(s) users are > running on the Windows side that causes this on our Linux Samba server. We have > hundreds of Samba clients and increasing the debug level for all does not > seem the way to go. I am also not sure as to which the suspect clients are > to zero in on them. > > Any thoughts or suggestions as to how to proceed with the debug, what log level to set, > and what strings to look for in (which) log files would be appreciated. > > Here are some details. > > uname -a > Linux plxs0131 2.4.9-45lxset36tcpsmp #1 SMP Wed Jan 12 09:16:34 PST 2005 i686 unknown > > smbd -V > Version 3.0.10 > > > Thanks, > > -KD-- Ben Walton Systems Programmer Office of Planning & IT Faculty of Arts & Science University of Toronto Cell: 416.407.5610 PGP Key Id: 8E89F6D2 -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 191 bytes Desc: This is a digitally signed message part Url : http://lists.samba.org/archive/samba/attachments/20060329/c375def4/attachment.bin