Samba 2.2.7a on RH 9.0
Installed, working partially
Three workstations seem to be working fine. A fourth one, with identical login fails when accessing, even for something as simple as ls /mnt/server1
The same error is repeated over and over, changing only generation # (pid stays same)
SMB server not responding
smb_get_length: recv error = 5
smb_trans2_request: result = -5, setting invalid
smb_retry: successful, new pid=<same number>, generation=<last_number++>
smb_proc_readdir_long: error=-5, retrying
and continues until ctrl-c'ed
Just tested this. The ls did as above. A small file copy worked (cp /mnt/server1/foo/bar gotit). Running a program and opening the same file r/w also worked. A cd /mnt/server1/foo also worked, as did cd /mnt/server1... but then ls while in that directory resulted in the error loop again.
Ooops - I figured it out; it turns out the other three w/s exhibit the same problem, but they don't go to the top level of the share at all, so it didn't show up until I did the same test from them. What's happening is that the mounted shared directory itself contained a dead mount to another machine. The problem is cured by unmounting the dead one.
But that brings up another question: is there a way to automagically unmount a share when it disappears (for whatever reason - hardware failure, software failure, wetware failure...)?