Bug 5222 - Bug.
Product: CifsVFS
Classification: Unclassified
Component: kernel fs
x86 Linux
: P3 regression
: ---
Assigned To: Steve French
Depends on:
  Show dependency treegraph
Reported: 2008-01-22 05:50 UTC by Mede
Modified: 2009-03-07 11:19 UTC (History)
1 user (show)

See Also:


Note You need to log in before you can comment on or make changes to this bug.
Description Mede 2008-01-22 05:50:39 UTC

My server on redhat AS 4 the machine hang when samba log this error :

Jan 22 10:45:17 s64p17bicu4 kernel:  CIFS VFS: No response buffer
Jan 22 10:45:17 s64p17bicu4 kernel:  CIFS VFS: Send error in Close = -9

Comment 1 Steve French 2008-08-28 19:59:11 UTC
What version of cifs vfs is this (RedHat has updated the version of cifs in their kernel updates a few times).   You can see the version by doing either
1) /sbin/modinfo of the file cifs.ko
2) "cat /proc/fs/cifs/DebugData"
Comment 2 Steve French 2008-12-05 14:03:26 UTC
The main cause of "-9" errors (Error bad file handle on close could sometimes be returned when the server or network hung, and the client reconnected but tried to close a file handle which has now been invalidated by the act of reconnecting the session).

It sounds like the more important issue in your case though is why the server hung on a particular command.

By the way, have you tried the updated kernels for RHEL (experimental even newer ones are available at http://people.redhat.com/jlayton)
Comment 3 Shirish S. Pargaonkar 2009-02-02 15:39:59 UTC
Is this still a problem?
Comment 4 Steve French 2009-03-07 11:19:13 UTC
Multiple fixes have been made in this area  1) to reduce or eliminate the "stuck sessions" by changing how we use the socket API in kernel  2) to properly invalidate file handles so we don't try to close file handles that are already closed after network crash