[Samba] Killing a process that has an open file on smbfs

Louis Lam lsauchun at hotmail.com
Wed Apr 3 17:59:03 GMT 2002


Hi,

I have Done the same test on a later kernel 2.4.7 found with redhat 7.2. The 
same thing occured, when killing a process that still has an open file on 
smbfs, the mountpoint seem unusable, even after the retry message appear on 
the console. Those error msgs are repeated.

Has this problem been fixed in later versions of the 2.4 kernel?

Here is the way to reproduce the problem.

1. Do smbmount a share from the server.

2. "vi" a file(or some other program that holds on to a file on the share)

3. kill the vi application(simulating non-graceful exit)

    Does it block?

If I try to "ls" that mountpoint or even try to smbumount, the process will
block.

    Does it cause more errors?


If you mean getting more of these messages:
smb_get_length: recv error = 512
smb_request: result -512 setting invalid
smb_retry: successful, new pid = xxxx, generation=xxx

These messages are repeated continuously.


    There should be a connection between localhost and port 139 on the 
server.
If you look at it with netstat it should be ESTABLISHED.


Indeed there was, when I do a netstat, I get the following:

On the client:

Proto Recv-Q Send-Q Local Address     Foreign Address      State
----- ------ ------ -------------     ---------------      ------
tcp   0      0      10.20.30.16:1025  kitchen:netbios-ssn  ESTABLISHED

On the Server:

Proto Recv-Q Send-Q Local Address        Foreign Address      State
----- ------ ------ -------------        ---------------      ------
tcp   0      0      kitchen:netbios-ssn  10.20.30.16:1024   ESTABLISHED


I don't see anything about port 139 though. Is that the netbios-ssn?



Thank you in advance.

Louis Lam


_________________________________________________________________
Send and receive Hotmail on your mobile device: http://mobile.msn.com





More information about the samba mailing list