Dragan Krnic
2007-Jul-16 13:08 UTC
[Samba] 3.0.25b-1.1.72-1411 - copy from and to the same samba share
A copy from a samba share to the same share frequently fails with error "The specified network name is no longer available". The remaining copy is the same size as source but corrupted. WinXP logs an obscure NetBT Event ID 4322, which says "NetBT could not process a request, because at least one OutOfResources-Exception occurred in the last hour". One can reproduce this error anytime if one chooses a big enough file and copies it a second time before the first copy is done. Samba 3.0.24 does not have this bug. When a double copy fails there is this line in full_audit log at the moment the errors pop up: Jul 16 14:52:15 p92 smbd_audit: [2007/07/16 14:52:15, 0] lib/util_sock.c:read_data(534) Jul 16 14:52:15 p92 smbd_audit: read_data: read failure for 4 bytes to client 172.24.204.176. Error = Connection reset by peer Is this a known problem ? Can I produce more diagnostics to help fix it ?
Jeremy Allison
2007-Jul-16 18:53 UTC
[Samba] 3.0.25b-1.1.72-1411 - copy from and to the same samba share
On Mon, Jul 16, 2007 at 03:08:01PM +0200, Dragan Krnic wrote:> A copy from a samba share to the same share frequently fails > with error "The specified network name is no longer available". > > The remaining copy is the same size as source but corrupted. > > WinXP logs an obscure NetBT Event ID 4322, which says > "NetBT could not process a request, because at least one > OutOfResources-Exception occurred in the last hour". > > One can reproduce this error anytime if one chooses a big enough > file and copies it a second time before the first copy is done. > Samba 3.0.24 does not have this bug. > > When a double copy fails there is this line in full_audit log > at the moment the errors pop up: > > Jul 16 14:52:15 p92 smbd_audit: [2007/07/16 14:52:15, 0] > lib/util_sock.c:read_data(534) > Jul 16 14:52:15 p92 smbd_audit: read_data: read failure for 4 bytes > to client 172.24.204.176. Error = Connection reset by peer > > Is this a known problem ? > Can I produce more diagnostics to help fix it ?Open a bug and attach an ethereal/wireshare network trace please. Jeremy.