We have serious problems with samba on our system and it seems like samba is the cause.
After the freeze, the only thing one can do is to reboot the machine via recovery console.
The setup (wich was not my idea) is as follows:
Client: Suse 9.3:
Server: Suse 9.3: only purpose: Samba-Server
The Client mounts shares consisting of directories full of images. The Client then serves these images trough apache to the web.
Examining /var/log/messages on the CLIENT machine
The log is shattered with:
kernel: smb_add_request: request [ffff810061ef16c0, mid=1581] timed out!
And directly before the freeze I could see:
kernel: smbiod_handle_request: smbiod got a request ... and we don't implement oplocks!
kernel: smb_receive_header: long packet: 65636
although the output directly before the freeze varied the last times, it always had something to do with samba.
Examining /var/log/messages on the SERVER machine
smbd: read_socket_data: recv failure for 4. Error = Connection reset by peer
which happens at the same time when smb_add_request time outs take place on the client machine.
Has anybody experienced similar problems? Can anybody see the cause for these freezes.
Does anybody know about performance problems with a webserver accessing files on samba shares?
If you need any more info on the system or more log output, I can provide that.
Thanks in advance!