Re: [Samba] Samba kills network.
Tested Denni's sugested fix no change. Striped my my config to the config below and the errors change but the end bug is the same: Apr 14 00:25:32 Ghost smbd[21811]: [2007/04/14 00:25:32, 0] lib/util_sock.c:write_data(562) Apr 14 00:25:32 Ghost smbd[21811]: write_data: write failure in writing to client 0.0.0.0. Error Connection reset by peer Apr 14 00:25:32 Ghost smbd[21811]: [2007/04/14 00:25:32, 0] lib/util_sock.c:send_smb(769) Apr 14 00:25:32 Ghost smbd[21811]: Error writing 4 bytes to client. -1. (Connection reset by peer) Apr 14 00:26:02 Ghost smbd[4063]: [2007/04/14 00:26:02, 0] smbd/oplock.c:oplock_timeout_handler(351) Apr 14 00:26:02 Ghost smbd[4063]: Oplock break failed for file Complete Torrents/Prison.Break.Complete.Season.1.DVDRip.torrent -- replying anyway Apr 14 00:26:03 Ghost smbd[21750]: [2007/04/14 00:26:03, 0] smbd/oplock.c:oplock_timeout_handler(351) Apr 14 00:26:03 Ghost smbd[21750]: Oplock break failed for file Auto Import/Prison.Break.Complete.Season.1.DVDRip.torrent -- replying anyway Apr 14 00:26:32 Ghost smbd[4063]: [2007/04/14 00:26:32, 0] smbd/oplock.c:oplock_timeout_handler(351) Apr 14 00:26:32 Ghost smbd[4063]: Oplock break failed for file Complete Torrents/Prison.Break.Complete.Season.1.DVDRip.torrent.part -- replying anyway Apr 14 00:29:05 Ghost nmbd[5690]: [2007/04/14 00:29:05, 0] nmbd/nmbd_incomingrequests.c:process_name_refresh_request(172) Apr 14 00:29:05 Ghost nmbd[5690]: process_name_refresh_request: unicast name registration request received for name VADER<20> from IP 192.168.1.97 on subnet UNICAST_SUBNET. Apr 14 00:29:05 Ghost nmbd[5690]: [2007/04/14 00:29:05, 0] nmbd/nmbd_incomingrequests.c:process_name_refresh_request(173) Apr 14 00:29:05 Ghost nmbd[5690]: Error - should be sent to WINS server Apr 14 00:29:05 Ghost nmbd[5690]: [2007/04/14 00:29:05, 0] nmbd/nmbd_incomingrequests.c:process_name_refresh_request(172) Apr 14 00:29:05 Ghost nmbd[5690]: process_name_refresh_request: unicast name registration request received for name FAMILY<00> from IP 192.168.1.97 on subnet UNICAST_SUBNET. Apr 14 00:29:05 Ghost nmbd[5690]: [2007/04/14 00:29:05, 0] nmbd/nmbd_incomingrequests.c:process_name_refresh_request(173) Apr 14 00:29:05 Ghost nmbd[5690]: Error - should be sent to WINS server Apr 14 00:29:05 Ghost nmbd[5690]: [2007/04/14 00:29:05, 0] nmbd/nmbd_incomingrequests.c:process_name_refresh_request(172) Apr 14 00:29:05 Ghost nmbd[5690]: process_name_refresh_request: unicast name registration request received for name VADER<00> from IP 192.168.1.97 on subnet UNICAST_SUBNET. Apr 14 00:29:05 Ghost nmbd[5690]: [2007/04/14 00:29:05, 0] nmbd/nmbd_incomingrequests.c:process_name_refresh_request(173) Apr 14 00:29:05 Ghost nmbd[5690]: Error - should be sent to WINS server Apr 14 00:29:20 Ghost smbd[21810]: [2007/04/14 00:29:20, 0] smbd/oplock.c:oplock_timeout_handler(351) Apr 14 00:29:20 Ghost smbd[21810]: Oplock break failed for file Prison.Break.Complete.Season.1.DVDRip.torrent -- replying anyway Apr 14 00:29:20 Ghost nmbd[5690]: [2007/04/14 00:29:20, 0] nmbd/nmbd_incomingrequests.c:process_name_refresh_request(172) Apr 14 00:29:20 Ghost nmbd[5690]: process_name_refresh_request: unicast name registration request received for name FAMILY<00> from IP 192.168.1.64 on subnet UNICAST_SUBNET. Apr 14 00:29:20 Ghost nmbd[5690]: [2007/04/14 00:29:20, 0] nmbd/nmbd_incomingrequests.c:process_name_refresh_request(173) Apr 14 00:29:20 Ghost nmbd[5690]: Error - should be sent to WINS server Apr 14 00:29:20 Ghost nmbd[5690]: [2007/04/14 00:29:20, 0] nmbd/nmbd_incomingrequests.c:process_name_refresh_request(172) Apr 14 00:29:20 Ghost nmbd[5690]: process_name_refresh_request: unicast name registration request received for name TIM<20> from IP 192.168.1.64 on subnet UNICAST_SUBNET. Apr 14 00:29:20 Ghost nmbd[5690]: [2007/04/14 00:29:20, 0] nmbd/nmbd_incomingrequests.c:process_name_refresh_request(173) Apr 14 00:29:20 Ghost nmbd[5690]: Error - should be sent to WINS server Apr 14 00:29:20 Ghost nmbd[5690]: [2007/04/14 00:29:20, 0] nmbd/nmbd_incomingrequests.c:process_name_refresh_request(172) Apr 14 00:29:20 Ghost nmbd[5690]: process_name_refresh_request: unicast name registration request received for name TIM<00> from IP 192.168.1.64 on subnet UNICAST_SUBNET. Apr 14 00:29:20 Ghost nmbd[5690]: [2007/04/14 00:29:20, 0] nmbd/nmbd_incomingrequests.c:process_name_refresh_request(173) Apr 14 00:29:20 Ghost nmbd[5690]: Error - should be sent to WINS server # This is the main Samba configuration file. You should read the # smb.conf(5) manual page in order to understand the options listed # here. Samba has a huge number of configurable options (perhaps too # many!) most of which are not shown in this example # # Any line which starts with a ; (semi-colon) or a # (hash) # is a comment and is ignored. In this example we will use a # # for commentry and a ; for parts of the config file that you # may wish to enable # # NOTE: Whe
Re: [Samba] Samba kills network.
Tom Peters wrote:
Nope Refreshing my shares does not help when i am getting errors on the
network. Still getting large amounts of errors generated and a hang of
all the computers using network processes:
Apr 11 01:08:37 Ghost smbd[32453]: [2007/04/11 01:08:37, 0]
lib/util_sock.c:write_data(562)
Apr 11 01:08:37 Ghost smbd[32453]: write_data: write failure in
writing to client 192.168.1.97. Error Broken pipe
Apr 11 01:08:37 Ghost smbd[32453]: [2007/04/11 01:08:37, 0]
lib/util_sock.c:send_smb(769)
Apr 11 01:08:37 Ghost smbd[32453]: Error writing 75 bytes to client.
-1. (Broken pipe)
Apr 11 01:09:18 Ghost smbd[9885]: [2007/04/11 01:09:18, 0]
lib/util_sock.c:read_data(534)
Apr 11 01:09:18 Ghost smbd[9885]: read_data: read failure for 4 bytes
to client 192.168.1.97. Error = Connection reset by peer
Apr 11 01:09:18 Ghost smbd[9885]: [2007/04/11 01:09:18, 0]
lib/util_sock.c:write_data(562)
Apr 11 01:09:18 Ghost smbd[9885]: write_data: write failure in writing
to client 192.168.1.97. Error Broken pipe
Apr 11 01:09:18 Ghost smbd[9885]: [2007/04/11 01:09:18, 0]
lib/util_sock.c:send_smb(769)
Apr 11 01:09:18 Ghost smbd[9885]: Error writing 75 bytes to client.
-1. (Broken pipe)
Tim Otten
At 10:14 AM 4/7/2007 +0200, C.Scheeder wrote:
Tom Peters schrieb:
By the way, I was hoping this would be significant: If I refresh my
view of a Samba share on the XP box by pressing F5 or clicking
View-->Refresh, so that there is some recent activity on that share,
and then begin the copy operation immediately, e.g. within 20
seconds, the copy operation always succeeds.
This makes me beleave it's a problem of your XP-installation.
Please give it a try to boot with a knoppix or another live-cd and
then test if the problems are reproduceable or not.
Roots <[EMAIL PROTECTED]> is that true for you as well?
The files I am copying tend to be medium-large: 180 mb to 900 mb.
But even with small files I've had it do the same thing.
At 10:00 AM 4/6/2007 +0200, C.Scheeder wrote:
Hi,
First:
you missed to tell us the version of samba you are using, and the
Software/OS of the client you are trying to copy your files from.
Yes, sorry, I realized after that we had not discussed versions or
etc. Of course I don't know what versions Roots <[EMAIL PROTECTED]>
but I have 3.0.9-2.3 of Samba here. I know it's pretty old.
and your kernelversion is?
2.6.8-24.25
second:
I never had any problems moving large files around using w98/W2k/XP
boxes to and from the samba-servers i maintain (roundabout 12
machines),
except when bad/dying hardware/software was involved.
have you tried from another client?
Yes. Similar issues result from another XP Pro box.
have you updated ALL your clients drivers (not only the Nic's drivers)
to the latest versions?
No.
Ok as you have tryend from two clients, that shouldn't be a problem,
as i guess they have diffrentz Hardware.
have you checked if there are corrupted packets on your network
("ifconfig eth1" shows you statistics off packets on your server)
It's eth0 for me. Out of 95 million packets received the errors
counts are: 1 error, 7 overruns, 1 frame, 0 dropped.
For 102.77 million packet sent, I have zero errors of any kind, zero
collisions, txqueue length 1000.
OK, that is bad. You should not get errors here with modern Hardware.
I checked the servers i can lay my hands on from here, and they have
no errors, nada. (one of them with over 300 million packets send and
received, and 5 other with lower counts, probably caused by wrapping
counters.)
have you replaced any peace of hardware inbetween the server and the
client?
No. Well, yes, but it made no difference. I swapped my SMC
10/100/1000 switch out for a 3Com SmartStack 10/100.
I did however try a crossover cable some time ago. The same results
(errors) occur.
Did you try and replace the NIC's in the clients and the server?
I have seen strange things happen with partly damaged/dying NIC's.
Even defekt onboard NIC's on brand new motherboards.
(if you don't have a spare-switch, connect client and server via an
crossover-cable...)
has your switch updateable firmware? if yes, have you updated it to
the
latest version?
I checked. There is no later version available for the SMC.
try to boot knoppix on the client and smbmount the shares then test
if then copying works.
Make shure your server is the masterbrowser at any time.
The os level is set to 33, which is supposed to make sure of that.
SUPPOSED, i've seen XP-computers steal the Masterbrowser-role even from
a samba server with os-level set to 255, took me half a Day to find the
machine and punish the user for having installed tcp-ip, netbios and
netware-protokoll on the machine.
make shure ALL unused network connections in your clients are
disabled. (like ieee1394-network, or dvb-network, unused
WLAN-Connetions and what the hell modern windows versions think of
devices being able to do networking)
Hmmm The 1394 connection was bou
RE: [Samba] Samba kills network.
All my other replies to this (5 of them so far) keep getting blocked by the SP*M filter. http://lists.samba.org/archive/samba/2007-February/129780.html Dennis -- To unsubscribe from this list go to the following URL and read the instructions: https://lists.samba.org/mailman/listinfo/samba
Re: [Samba] Samba kills network.
At 10:14 AM 4/7/2007 +0200, C.Scheeder wrote:
Tom Peters schrieb:
By the way, I was hoping this would be significant: If I refresh my view
of a Samba share on the XP box by pressing F5 or clicking View-->Refresh,
so that there is some recent activity on that share, and then begin the
copy operation immediately, e.g. within 20 seconds, the copy operation
always succeeds.
This makes me beleave it's a problem of your XP-installation.
Please give it a try to boot with a knoppix or another live-cd and then
test if the problems are reproduceable or not.
Roots <[EMAIL PROTECTED]> is that true for you as well?
The files I am copying tend to be medium-large: 180 mb to 900 mb. But
even with small files I've had it do the same thing.
At 10:00 AM 4/6/2007 +0200, C.Scheeder wrote:
Hi,
First:
you missed to tell us the version of samba you are using, and the
Software/OS of the client you are trying to copy your files from.
Yes, sorry, I realized after that we had not discussed versions or etc.
Of course I don't know what versions Roots <[EMAIL PROTECTED]> but I have
3.0.9-2.3 of Samba here. I know it's pretty old.
and your kernelversion is?
2.6.8-24.25
second:
I never had any problems moving large files around using w98/W2k/XP
boxes to and from the samba-servers i maintain (roundabout 12 machines),
except when bad/dying hardware/software was involved.
have you tried from another client?
Yes. Similar issues result from another XP Pro box.
have you updated ALL your clients drivers (not only the Nic's drivers)
to the latest versions?
No.
Ok as you have tryend from two clients, that shouldn't be a problem,
as i guess they have diffrentz Hardware.
have you checked if there are corrupted packets on your network
("ifconfig eth1" shows you statistics off packets on your server)
It's eth0 for me. Out of 95 million packets received the errors counts
are: 1 error, 7 overruns, 1 frame, 0 dropped.
For 102.77 million packet sent, I have zero errors of any kind, zero
collisions, txqueue length 1000.
OK, that is bad. You should not get errors here with modern Hardware.
I checked the servers i can lay my hands on from here, and they have no
errors, nada. (one of them with over 300 million packets send and
received, and 5 other with lower counts, probably caused by wrapping counters.)
have you replaced any peace of hardware inbetween the server and the
client?
No. Well, yes, but it made no difference. I swapped my SMC 10/100/1000
switch out for a 3Com SmartStack 10/100.
I did however try a crossover cable some time ago. The same results
(errors) occur.
Did you try and replace the NIC's in the clients and the server?
I have seen strange things happen with partly damaged/dying NIC's.
Even defekt onboard NIC's on brand new motherboards.
(if you don't have a spare-switch, connect client and server via an
crossover-cable...)
has your switch updateable firmware? if yes, have you updated it to the
latest version?
I checked. There is no later version available for the SMC.
try to boot knoppix on the client and smbmount the shares then test if
then copying works.
Make shure your server is the masterbrowser at any time.
The os level is set to 33, which is supposed to make sure of that.
SUPPOSED, i've seen XP-computers steal the Masterbrowser-role even from
a samba server with os-level set to 255, took me half a Day to find the
machine and punish the user for having installed tcp-ip, netbios and
netware-protokoll on the machine.
make shure ALL unused network connections in your clients are disabled.
(like ieee1394-network, or dvb-network, unused WLAN-Connetions and what
the hell modern windows versions think of devices being able to do networking)
Hmmm The 1394 connection was bound to Client for Microsoft Windows,
Service Advertising Protocol, File and Printer Sharing for MS Networks, and
Internet protocol.
I don't anticipate using the 1394 connection for anything soon, so I
cleared the checkboxes for most of that. SAP is greyed-out, can't change
that one.
My first attempt to copy files to a samba drive failed immediately, just as
before. Perhaps a therapeutic reboot is necessary.
-Tom
Christoph
Christoph
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba
-
933. [Haiku error messages] Wind catches lily / scattering petals to the wind:
/ segmentation fault
--... ...-- -.. . -. . --.- --.- -...
[EMAIL PROTECTED] (remove "nospam") N9QQB (amateur radio)
"HEY YOU" (loud shouting)WEB: http://www.mixweb.com/tpeters
43° 7' 17.2" N by 88° 6' 28.9" W, Elevation 815', Grid Square EN53wc
WAN/LAN/Telcom Analyst, Tech Writer, MCP, CCNA, Registered Linux User 385531
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba
Re: [Samba] Samba kills network.
Roots schrieb: Client OS: Windows XP SP2, Samba 3.0.24-3.fc6. Tested from another client: Yes i have more than one computer that both have the same problems. Updated NIC Drivers: I have checked for updated drivers for clients but have found non. I im unsure what driver version is built into the kernel on the server the NIC is e1000 (82545GM). So No. Corrupt Packets: I had some problems with this but replaced some cables these problems have gone away but samba problems continue. Replaced Hardware: Nope, I have no problems with file swapping between XP clients. so at least please try to connect your server to a diffrent port of your switch, i have seen single ports of switches die. Updated Switch Firmware: My switch is unmanaged (Netgear GS116). Booting from a live CD: I booted from knoppix and dint appear to have any problems. However as its an intermittent problem i cant be sure. UNC or Mapped drive: I'm using mapped drives. Make sure server is master browsers: Yes it is. The Problem: I find the problem main occurs when i try to swap file between network drivers on the server from the. Or when dumping files on the server, or when saving files to the server from firefox. I don't find the file size makes any difference. Tom, in answer to your question im having trouble testing if your fix works for me as i'm not seeing the problem at the moment all though I'm sure it will return. Ive included a copy of my config, the errors in a text file and a ethereal cap of my network during one of the errors. If someone with a better understand of network protocols than me was to take a look at it take the trouble to look at the cap im sure they would be able to figure out whats going on. All the files are in this rar: http://timo.bulldoghome.com/pages/timo_bulldoghome_com/samba.rar ok, i have a few questions/comments about your smb.conf: first: keep it as simple as possible, don't set options you do not need to change from the default. [global] getwd cache = yes socket options = TCP_NODELAY SO_SNDBUF=65536 SO_RCVBUF=65536 IPTOS_LOWDELAY i guess you have a recent kernel (2.6.xx), so get rid of these SO_SNDBUF/SO_RCBUFF options. There has been a discussion on these options on the list a few Weeks ago, and the conclusion was that they are bad more or less withe recent kernels. use sendfile = no why do you disable sendfile here, if you later on enable it for most of your shares? lock spin time = 15 lock spin count = 30 "lock spin count" is non-functional with your samba-version, and "lock spin time" has a default of 200 microseconds, why do you reduce it to 15 microseconds? map to guest = bad user log level = 1 security = user os level = 64 local master = Yes time server = Yes domain master = yes preferred master = yes wins support = yes domain logons = yes dos filetimes = Yes workgroup = Family netbios name = Ghost server string = Samba Server %v printcap name = cups load printers = yes printing = cups printer admin = root @adm @is log file = /var/log/samba/log.%m max log size = 50 ;hosts allow = 192.168.1.0/24 127.0.0.1 interfaces = eth1 lo bind interfaces only = yes encrypt passwords = yes smb passwd file = /etc/samba/smbpasswd add user script = /usr/sbin/useradd -d /dev/null -g machines -c 'Machine Account' -s /bin/false -M %u wins proxy = yes you do not need "wins-proxy" with modern clients (w2k and newer) dns proxy = no logon path = logon drive = U: logon script = %U.bat oplocks = no level2 oplocks = no why do you disable oplocks here, if you later on enable it for most of your shares? change notify timeout = 300 lpq cache time = 30 winbind uid = 1-2 winbind gid = 1-2 winbind separator = + you don't want winbind-stuff at all, as you are the domain-controler. please remove these options. (and don't start winbind, you only need it if you are not the PDC of the domain your authenticating agains) oplocks = no level2 oplocks = no Huh, why disable oplocks again?!? ; deadtime = 60 wins proxy = yes lpq cache time = 30 change notify timeout = 300 getwd cache = yes dos filetimes = yes domain logons = yes obey pam restrictions = yes unix password sync = Yes pam password change = yes passwd program = /usr/bin/passwd %u passwd chat = *New*UNIX*password* %n\n *Re*ype*new*UNIX*password* %n\n \ *passwd:*all*authentication*tokens*updated*successfully* add user script = /usr/sbin/useradd -s /bin/false '%u' delete user script = /usr/sbin/userdel '%s' add user to group script =
Re: [Samba] Samba kills network.
Tom Peters schrieb:
By the way, I was hoping this would be significant: If I refresh my view
of a Samba share on the XP box by pressing F5 or clicking
View-->Refresh, so that there is some recent activity on that share, and
then begin the copy operation immediately, e.g. within 20 seconds, the
copy operation always succeeds.
This makes me beleave it's a problem of your XP-installation.
Please give it a try to boot with a knoppix or another live-cd and then
test if the problems are reproduceable or not.
Roots <[EMAIL PROTECTED]> is that true for you as well?
The files I am copying tend to be medium-large: 180 mb to 900 mb. But
even with small files I've had it do the same thing.
At 10:00 AM 4/6/2007 +0200, C.Scheeder wrote:
Hi,
First:
you missed to tell us the version of samba you are using, and the
Software/OS of the client you are trying to copy your files from.
Yes, sorry, I realized after that we had not discussed versions or etc.
Of course I don't know what versions Roots <[EMAIL PROTECTED]> but I have
3.0.9-2.3 of Samba here. I know it's pretty old.
and your kernelversion is?
second:
I never had any problems moving large files around using w98/W2k/XP
boxes to and from the samba-servers i maintain (roundabout 12 machines),
except when bad/dying hardware/software was involved.
have you tried from another client?
Yes. Similar issues result from another XP Pro box.
have you updated ALL your clients drivers (not only the Nic's drivers)
to the latest versions?
No.
Ok as you have tryend from two clients, that shouldn't be a problem,
as i guess they have diffrentz Hardware.
have you checked if there are corrupted packets on your network
("ifconfig eth1" shows you statistics off packets on your server)
It's eth0 for me. Out of 95 million packets received the errors counts
are: 1 error, 7 overruns, 1 frame, 0 dropped.
For 102.77 million packet sent, I have zero errors of any kind, zero
collisions, txqueue length 1000.
OK, that is bad. You should not get errors here with modern Hardware.
I checked the servers i can lay my hands on from here, and they have no
errors, nada. (one of them with over 300 million packets send and
received, and 5 other with lower counts, probably caused by wrapping
counters.)
have you replaced any peace of hardware inbetween the server and the
client?
No. Well, yes, but it made no difference. I swapped my SMC 10/100/1000
switch out for a 3Com SmartStack 10/100.
I did however try a crossover cable some time ago. The same results
(errors) occur.
Did you try and replace the NIC's in the clients and the server?
I have seen strange things happen with partly damaged/dying NIC's.
Even defekt onboard NIC's on brand new motherboards.
(if you don't have a spare-switch, connect client and server via an
crossover-cable...)
has your switch updateable firmware? if yes, have you updated it to the
latest version?
I checked. There is no later version available for the SMC.
try to boot knoppix on the client and smbmount the shares then test if
then copying works.
Make shure your server is the masterbrowser at any time.
The os level is set to 33, which is supposed to make sure of that.
SUPPOSED, i've seen XP-computers steal the Masterbrowser-role even from
a samba server with os-level set to 255, took me half a Day to find the
machine and punish the user for having installed tcp-ip, netbios and
netware-protokoll on the machine.
make shure ALL unused network connections in your clients are
disabled. (like ieee1394-network, or dvb-network, unused
WLAN-Connetions and what the hell modern windows versions think of
devices being able to do networking)
Christoph
Christoph
--
To unsubscribe from this list go to the following URL and read the
instructions: https://lists.samba.org/mailman/listinfo/samba
Re: [Samba] Samba kills network.
Client OS: Windows XP SP2, Samba 3.0.24-3.fc6. Tested from another client: Yes i have more than one computer that both have the same problems. Updated NIC Drivers: I have checked for updated drivers for clients but have found non. I im unsure what driver version is built into the kernel on the server the NIC is e1000 (82545GM). So No. Corrupt Packets: I had some problems with this but replaced some cables these problems have gone away but samba problems continue. Replaced Hardware: Nope, I have no problems with file swapping between XP clients. Updated Switch Firmware: My switch is unmanaged (Netgear GS116). Booting from a live CD: I booted from knoppix and dint appear to have any problems. However as its an intermittent problem i cant be sure. UNC or Mapped drive: I'm using mapped drives. Make sure server is master browsers: Yes it is. The Problem: I find the problem main occurs when i try to swap file between network drivers on the server from the. Or when dumping files on the server, or when saving files to the server from firefox. I don't find the file size makes any difference. Tom, in answer to your question im having trouble testing if your fix works for me as i'm not seeing the problem at the moment all though I'm sure it will return. Ive included a copy of my config, the errors in a text file and a ethereal cap of my network during one of the errors. If someone with a better understand of network protocols than me was to take a look at it take the trouble to look at the cap im sure they would be able to figure out whats going on. All the files are in this rar: http://timo.bulldoghome.com/pages/timo_bulldoghome_com/samba.rar Tim Otten -- To unsubscribe from this list go to the following URL and read the instructions: https://lists.samba.org/mailman/listinfo/samba
Re: [Samba] Samba kills network.
By the way, I was hoping this would be significant: If I refresh my view of
a Samba share on the XP box by pressing F5 or clicking View-->Refresh, so
that there is some recent activity on that share, and then begin the copy
operation immediately, e.g. within 20 seconds, the copy operation always
succeeds.
Roots <[EMAIL PROTECTED]> is that true for you as well?
The files I am copying tend to be medium-large: 180 mb to 900 mb. But even
with small files I've had it do the same thing.
At 10:00 AM 4/6/2007 +0200, C.Scheeder wrote:
Hi,
First:
you missed to tell us the version of samba you are using, and the
Software/OS of the client you are trying to copy your files from.
Yes, sorry, I realized after that we had not discussed versions or etc. Of
course I don't know what versions Roots <[EMAIL PROTECTED]> but I have
3.0.9-2.3 of Samba here. I know it's pretty old.
second:
I never had any problems moving large files around using w98/W2k/XP
boxes to and from the samba-servers i maintain (roundabout 12 machines),
except when bad/dying hardware/software was involved.
have you tried from another client?
Yes. Similar issues result from another XP Pro box.
have you updated ALL your clients drivers (not only the Nic's drivers)
to the latest versions?
No.
have you checked if there are corrupted packets on your network ("ifconfig
eth1" shows you statistics off packets on your server)
It's eth0 for me. Out of 95 million packets received the errors counts are:
1 error, 7 overruns, 1 frame, 0 dropped.
For 102.77 million packet sent, I have zero errors of any kind, zero
collisions, txqueue length 1000.
have you replaced any peace of hardware inbetween the server and the
client?
No. Well, yes, but it made no difference. I swapped my SMC 10/100/1000
switch out for a 3Com SmartStack 10/100.
I did however try a crossover cable some time ago. The same results
(errors) occur.
(if you don't have a spare-switch, connect client and server via an
crossover-cable...)
has your switch updateable firmware? if yes, have you updated it to the
latest version?
I checked. There is no later version available for the SMC.
try to boot knoppix on the client and smbmount the shares then test if
then copying works.
Make shure your server is the masterbrowser at any time.
The os level is set to 33, which is supposed to make sure of that.
make shure ALL unused network connections in your clients are disabled.
(like ieee1394-network, or dvb-network, unused WLAN-Connetions and what
the hell modern windows versions think of devices being able to do networking)
Christoph
Tom Peters schrieb:
At 09:14 PM 4/5/2007 +0100, you wrote:
When i try to swap files between disks on my server using samba i get
errors: then windows reports "The specified network name is no longer
available". Sometimes the copies are fine other times files just wont
copy. Hardware checks out fine. Ive even replaced the network cable
between server and switch thinking that might be causing the problem.
Ive included a ethereal cap of the network during one of these problems,
and my samba config as attachments the errors are as follows:
Apr 5 20:41:53 Ghost smbd[27010]: [2007/04/05 20:41:53, 0]
lib/util_sock.c:get_peer_addr(1229)
Apr 5 20:41:53 Ghost smbd[27010]: getpeername failed. Error was
Transport endpoint is not connected
Please Help im at a loss to what the problem is!
[global]
getwd cache = yes
socket options = TCP_NODELAY SO_SNDBUF=65536 SO_RCVBUF=65536
IPTOS_LOWDELAY
use sendfile = no
lock spin time = 15
lock spin count = 30
map to guest = bad user
log level = 1
security = user
os level = 64
local master = Yes
time server = Yes
domain master = yes
preferred master = yes
wins support = yes
domain logons = yes
dos filetimes = Yes
workgroup = Family
netbios name = fileserver
server string = Samba Server %v
printcap name = cups
load printers = yes
printing = cups
printer admin = root @adm @is
log file = /var/log/samba/log.%m
max log size = 50
;hosts allow = 192.168.1.0/24 127.0.0.1
interfaces = eth1 lo
bind interfaces only = yes
encrypt passwords = yes
smb passwd file = /etc/samba/smbpasswd
add user script = /usr/sbin/useradd -d /dev/null -g machines -c
'Machine Account' -s /bin/false -M %u
wins proxy = yes
dns proxy = no
logon path =
logon drive = U:
logon script = %U.bat
oplocks = no
level2 oplocks = no
change notify timeout = 300
lpq cache time = 30
winbind uid = 1-2
winbind gid = 1-2
winbind separator = +
oplocks = no
level2 oplocks = no
; deadtime = 60
wins proxy = yes
lpq cache time = 30
change notify
Re: [Samba] Samba kills network.
Hi,
First:
you missed to tell us the version of samba you are using, and the
Software/OS of the client you are trying to copy your files from.
second:
I never had any problems moving large files around using w98/W2k/XP
boxes to and from the samba-servers i maintain (roundabout 12 machines),
except when bad/dying hardware/software was involved.
have you tried from another client?
have you updated ALL your clients drivers (not only the Nic's drivers)
to the latest versions?
have you checked if there are corrupted packets on your network
("ifconfig eth1" shows you statistics off packets on your server)
have you replaced any peace of hardware inbetween the server and the
client?
(if you don't have a spare-switch, connect client and server via an
crossover-cable...)
has your switch updateable firmware? if yes, have you updated it to the
latest version?
try to boot knoppix on the client and smbmount the shares then test if
then copying works.
Make shure your server is the masterbrowser at any time.
make shure ALL unused network connections in your clients are disabled.
(like ieee1394-network, or dvb-network, unused WLAN-Connetions and what
the hell modern windows versions think of devices being able to do
networking)
Christoph
Tom Peters schrieb:
At 09:14 PM 4/5/2007 +0100, you wrote:
When i try to swap files between disks on my server using samba i get
errors: then windows reports "The specified network name is no longer
available". Sometimes the copies are fine other times files just wont
copy. Hardware checks out fine. Ive even replaced the network cable
between server and switch thinking that might be causing the problem.
Ive included a ethereal cap of the network during one of these
problems, and my samba config as attachments the errors are as follows:
Apr 5 20:41:53 Ghost smbd[27010]: [2007/04/05 20:41:53, 0]
lib/util_sock.c:get_peer_addr(1229)
Apr 5 20:41:53 Ghost smbd[27010]: getpeername failed. Error was
Transport endpoint is not connected
Please Help im at a loss to what the problem is!
[global]
getwd cache = yes
socket options = TCP_NODELAY SO_SNDBUF=65536 SO_RCVBUF=65536
IPTOS_LOWDELAY
use sendfile = no
lock spin time = 15
lock spin count = 30
map to guest = bad user
log level = 1
security = user
os level = 64
local master = Yes
time server = Yes
domain master = yes
preferred master = yes
wins support = yes
domain logons = yes
dos filetimes = Yes
workgroup = Family
netbios name = fileserver
server string = Samba Server %v
printcap name = cups
load printers = yes
printing = cups
printer admin = root @adm @is
log file = /var/log/samba/log.%m
max log size = 50
;hosts allow = 192.168.1.0/24 127.0.0.1
interfaces = eth1 lo
bind interfaces only = yes
encrypt passwords = yes
smb passwd file = /etc/samba/smbpasswd
add user script = /usr/sbin/useradd -d /dev/null -g machines
-c 'Machine Account' -s /bin/false -M %u
wins proxy = yes
dns proxy = no
logon path =
logon drive = U:
logon script = %U.bat
oplocks = no
level2 oplocks = no
change notify timeout = 300
lpq cache time = 30
winbind uid = 1-2
winbind gid = 1-2
winbind separator = +
oplocks = no
level2 oplocks = no
; deadtime = 60
wins proxy = yes
lpq cache time = 30
change notify timeout = 300
getwd cache = yes
dos filetimes = yes
domain logons = yes
obey pam restrictions = yes
unix password sync = Yes
pam password change = yes
passwd program = /usr/bin/passwd %u
passwd chat = *New*UNIX*password* %n\n
*Re*ype*new*UNIX*password* %n\n \
*passwd:*all*authentication*tokens*updated*successfully*
add user script = /usr/sbin/useradd -s /bin/false '%u'
delete user script = /usr/sbin/userdel '%s'
add user to group script = /usr/bin/gpasswd -a '%u' '%g'
delete user from group script = /usr/bin/gpasswd -d '%u' '%g'
set primary group script = /usr/sbin/usermod -g '%g' '%u'
add group script = /usr/sbin/groupadd %g && getent group
'%g'|awk -F: '{print $3}'
delete group script = /usr/sbin/groupdel '%g'
add machine script = /usr/sbin/useradd -d /dev/null -g
machines -c 'Machine Account' -s /bin/false -M %u
[homes]
comment = Home Directories
browseable = no
writable = yes
oplocks = yes
level2 oplocks = yes
; use sendfile = yes
; preexec = echo "%u, %G, %a, %m (%I)\" >>/tmp/.log
[netlogon]
comment = Network Logon Service
path = /var/lib/samba/netlogon
writable = no
read only = yes
guest ok = no
Re: [Samba] Samba kills network.
Tom Peters wrote: At 03:04 PM 4/5/2007 -0700, Jeremy Allison wrote: On Thu, Apr 05, 2007 at 04:15:33PM -0500, Tom Peters wrote: > > I have this problem all the time. Samba for me will cause the "The > specified network name no longer exists" (that's how it's worded for me, > not "available") and copy a zero-length file to the destination drive. > > Reads from a Samba share are never any problem for me. > > If you immediately retry the operation, and immediately say "Y" to > overwrite it, it will succeed. > > If you do the above for the first in a series of files to be copied to a > Samba share, it will copy all the rest of them sucessfully. > > The other thing that gives this error away is the "getpeername > failed... ...transport endpoint not connected. " > > I thought I had this fixed, but after months, it has recurred. The fix I > tried was this (in smb.conf): > smb ports = 139 > > The explanation I got was that Windows try to connect to a server over > ports 443 and 139 nearly simultaneously, then use whichever one responds > first. Samba replies to both, and it might be that Windows has already > decided which it's going to use, and interprets the double reply as a > failure. > > Use sendfile = no has also been suggested to me. > > Frankly, this is embarrassing, and has kept me from pushing Samba harder. > When I ask about it, nobody seems to have a real answer. Easy enough to fix. Add : smb ports = 445 to the [global] section of your smb.conf. Pre-Windows 2000 clients won't be able to connect though. I tried that. Used 445, not 443 (oops). Did a load/unload of smbd. No change. Changed it to 139. Worked better, for a few months. Went back to old behavior, which is a lot like having no "smb ports =" in my conf file at all. Confirmed for me also add "smb ports = 445" to config file does not help. Tim -- To unsubscribe from this list go to the following URL and read the instructions: https://lists.samba.org/mailman/listinfo/samba
Re: [Samba] Samba kills network.
At 03:04 PM 4/5/2007 -0700, Jeremy Allison wrote: On Thu, Apr 05, 2007 at 04:15:33PM -0500, Tom Peters wrote: > > I have this problem all the time. Samba for me will cause the "The > specified network name no longer exists" (that's how it's worded for me, > not "available") and copy a zero-length file to the destination drive. > > Reads from a Samba share are never any problem for me. > > If you immediately retry the operation, and immediately say "Y" to > overwrite it, it will succeed. > > If you do the above for the first in a series of files to be copied to a > Samba share, it will copy all the rest of them sucessfully. > > The other thing that gives this error away is the "getpeername > failed... ...transport endpoint not connected. " > > I thought I had this fixed, but after months, it has recurred. The fix I > tried was this (in smb.conf): > smb ports = 139 > > The explanation I got was that Windows try to connect to a server over > ports 443 and 139 nearly simultaneously, then use whichever one responds > first. Samba replies to both, and it might be that Windows has already > decided which it's going to use, and interprets the double reply as a > failure. > > Use sendfile = no has also been suggested to me. > > Frankly, this is embarrassing, and has kept me from pushing Samba harder. > When I ask about it, nobody seems to have a real answer. Easy enough to fix. Add : smb ports = 445 to the [global] section of your smb.conf. Pre-Windows 2000 clients won't be able to connect though. I tried that. Used 445, not 443 (oops). Did a load/unload of smbd. No change. Changed it to 139. Worked better, for a few months. Went back to old behavior, which is a lot like having no "smb ports =" in my conf file at all. - 844. My veal cutlet tried to beat the s**t out of my cup of coffee... the coffee just wasn't strong enough to defend himself. (Tom Waits) --... ...-- -.. . -. . --.- --.- -... [EMAIL PROTECTED] (remove "nospam") N9QQB (amateur radio) "HEY YOU" (loud shouting)WEB: http://www.mixweb.com/tpeters 43° 7' 17.2" N by 88° 6' 28.9" W, Elevation 815', Grid Square EN53wc WAN/LAN/Telcom Analyst, Tech Writer, MCP, CCNA, Registered Linux User 385531 -- To unsubscribe from this list go to the following URL and read the instructions: https://lists.samba.org/mailman/listinfo/samba
Re: [Samba] Samba kills network.
On Thu, Apr 05, 2007 at 04:15:33PM -0500, Tom Peters wrote: > > I have this problem all the time. Samba for me will cause the "The > specified network name no longer exists" (that's how it's worded for me, > not "available") and copy a zero-length file to the destination drive. > > Reads from a Samba share are never any problem for me. > > If you immediately retry the operation, and immediately say "Y" to > overwrite it, it will succeed. > > If you do the above for the first in a series of files to be copied to a > Samba share, it will copy all the rest of them sucessfully. > > The other thing that gives this error away is the "getpeername > failed... ...transport endpoint not connected. " > > I thought I had this fixed, but after months, it has recurred. The fix I > tried was this (in smb.conf): > smb ports = 139 > > The explanation I got was that Windows try to connect to a server over > ports 443 and 139 nearly simultaneously, then use whichever one responds > first. Samba replies to both, and it might be that Windows has already > decided which it's going to use, and interprets the double reply as a > failure. > > Use sendfile = no has also been suggested to me. > > Frankly, this is embarrassing, and has kept me from pushing Samba harder. > When I ask about it, nobody seems to have a real answer. Easy enough to fix. Add : smb ports = 445 to the [global] section of your smb.conf. Pre-Windows 2000 clients won't be able to connect though. Jeremy. -- To unsubscribe from this list go to the following URL and read the instructions: https://lists.samba.org/mailman/listinfo/samba
Re: [Samba] Samba kills network.
At 09:14 PM 4/5/2007 +0100, you wrote:
When i try to swap files between disks on my server using samba i get
errors: then windows reports "The specified network name is no longer
available". Sometimes the copies are fine other times files just wont
copy. Hardware checks out fine. Ive even replaced the network cable
between server and switch thinking that might be causing the problem. Ive
included a ethereal cap of the network during one of these problems, and
my samba config as attachments the errors are as follows:
Apr 5 20:41:53 Ghost smbd[27010]: [2007/04/05 20:41:53, 0]
lib/util_sock.c:get_peer_addr(1229)
Apr 5 20:41:53 Ghost smbd[27010]: getpeername failed. Error was
Transport endpoint is not connected
Please Help im at a loss to what the problem is!
[global]
getwd cache = yes
socket options = TCP_NODELAY SO_SNDBUF=65536 SO_RCVBUF=65536
IPTOS_LOWDELAY
use sendfile = no
lock spin time = 15
lock spin count = 30
map to guest = bad user
log level = 1
security = user
os level = 64
local master = Yes
time server = Yes
domain master = yes
preferred master = yes
wins support = yes
domain logons = yes
dos filetimes = Yes
workgroup = Family
netbios name = fileserver
server string = Samba Server %v
printcap name = cups
load printers = yes
printing = cups
printer admin = root @adm @is
log file = /var/log/samba/log.%m
max log size = 50
;hosts allow = 192.168.1.0/24 127.0.0.1
interfaces = eth1 lo
bind interfaces only = yes
encrypt passwords = yes
smb passwd file = /etc/samba/smbpasswd
add user script = /usr/sbin/useradd -d /dev/null -g machines -c
'Machine Account' -s /bin/false -M %u
wins proxy = yes
dns proxy = no
logon path =
logon drive = U:
logon script = %U.bat
oplocks = no
level2 oplocks = no
change notify timeout = 300
lpq cache time = 30
winbind uid = 1-2
winbind gid = 1-2
winbind separator = +
oplocks = no
level2 oplocks = no
; deadtime = 60
wins proxy = yes
lpq cache time = 30
change notify timeout = 300
getwd cache = yes
dos filetimes = yes
domain logons = yes
obey pam restrictions = yes
unix password sync = Yes
pam password change = yes
passwd program = /usr/bin/passwd %u
passwd chat = *New*UNIX*password* %n\n *Re*ype*new*UNIX*password*
%n\n \
*passwd:*all*authentication*tokens*updated*successfully*
add user script = /usr/sbin/useradd -s /bin/false '%u'
delete user script = /usr/sbin/userdel '%s'
add user to group script = /usr/bin/gpasswd -a '%u' '%g'
delete user from group script = /usr/bin/gpasswd -d '%u' '%g'
set primary group script = /usr/sbin/usermod -g '%g' '%u'
add group script = /usr/sbin/groupadd %g && getent group '%g'|awk
-F: '{print $3}'
delete group script = /usr/sbin/groupdel '%g'
add machine script = /usr/sbin/useradd -d /dev/null -g machines
-c 'Machine Account' -s /bin/false -M %u
[homes]
comment = Home Directories
browseable = no
writable = yes
oplocks = yes
level2 oplocks = yes
; use sendfile = yes
; preexec = echo "%u, %G, %a, %m (%I)\" >>/tmp/.log
[netlogon]
comment = Network Logon Service
path = /var/lib/samba/netlogon
writable = no
read only = yes
guest ok = no
browseable = no
share modes = no
root preexec = /usr/bin/ntlogon -u %U -g %G -o %a -d
/var/lib/samba/netlogon/
root postexec = rm -f /var/lib/samba/netlogon/%U.bat
[www]
use sendfile = yes
comment = Web Site
path = /var/www
writeable = yes
valid users = @adm
force group = apache
force user = root
create mask = 3754
directory mask = 3754
force create mode = 3754
force directory mode = 3754
oplocks = Yes
level2 oplocks = yes
[logs]
comment = Server Log Files
path = /var/log
read only = yes
force group = root
force user = root
public = no
valid users = @adm
[etc]
comment = Server ETC Files
path = /etc
read only = yes
force group = root
force user = root
public = no
valid users = @adm
[Video's]
comment = Video's
path = /disks/sda
public = yes
only guest = yes
writable = yes
printable = no
oplocks = Yes
level2 oplocks = yes
use sendfile = yes
[Games]
comment = Games
path = /disks/hde
public = yes
only guest = yes
[Samba] Samba kills network.
When i try to swap files between disks on my server using samba i get
errors: then windows reports "The specified network name is no longer
available". Sometimes the copies are fine other times files just wont
copy. Hardware checks out fine. Ive even replaced the network cable
between server and switch thinking that might be causing the problem.
Ive included a ethereal cap of the network during one of these problems,
and my samba config as attachments the errors are as follows:
Apr 5 20:41:53 Ghost smbd[27010]: [2007/04/05 20:41:53, 0]
lib/util_sock.c:get_peer_addr(1229)
Apr 5 20:41:53 Ghost smbd[27010]: getpeername failed. Error was
Transport endpoint is not connected
Apr 5 20:41:53 Ghost smbd[32631]: [2007/04/05 20:41:53, 0]
lib/util_sock.c:get_peer_addr(1229)
Apr 5 20:41:53 Ghost smbd[32631]: getpeername failed. Error was
Transport endpoint is not connected
Apr 5 20:41:53 Ghost smbd[32631]: [2007/04/05 20:41:53, 0]
lib/util_sock.c:write_data(562)
Apr 5 20:41:53 Ghost smbd[32631]: write_data: write failure in
writing to client 0.0.0.0. Error Connection reset by peer
Apr 5 20:41:53 Ghost smbd[32631]: [2007/04/05 20:41:53, 0]
lib/util_sock.c:send_smb(769)
Apr 5 20:41:53 Ghost smbd[32631]: Error writing 4 bytes to client.
-1. (Connection reset by peer)
Apr 5 20:43:45 Ghost smbd[32650]: [2007/04/05 20:43:45, 0]
lib/util_sock.c:read_data(534)
Apr 5 20:43:45 Ghost smbd[32650]: read_data: read failure for 4 bytes
to client 192.168.1.97. Error = Connection reset by peer
Apr 5 20:43:45 Ghost smbd[32650]: [2007/04/05 20:43:45, 0]
lib/util_sock.c:write_data(562)
Apr 5 20:43:45 Ghost smbd[32650]: write_data: write failure in
writing to client 192.168.1.97. Error Broken pipe
Apr 5 20:43:45 Ghost smbd[32650]: [2007/04/05 20:43:45, 0]
lib/util_sock.c:send_smb(769)
Apr 5 20:43:45 Ghost smbd[32650]: Error writing 75 bytes to client.
-1. (Broken pipe)
Apr 5 20:44:16 Ghost smbd[32705]: [2007/04/05 20:44:16, 0]
lib/util_sock.c:get_peer_addr(1229)
Apr 5 20:44:16 Ghost smbd[32705]: getpeername failed. Error was
Transport endpoint is not connected
Apr 5 20:44:16 Ghost smbd[32705]: [2007/04/05 20:44:16, 0]
lib/util_sock.c:write_data(562)
Apr 5 20:44:16 Ghost smbd[32705]: write_data: write failure in
writing to client 192.168.1.97. Error Connection reset by peer
Apr 5 20:44:16 Ghost smbd[32705]: [2007/04/05 20:44:16, 0]
lib/util_sock.c:send_smb(769)
Apr 5 20:44:16 Ghost smbd[32705]: Error writing 4 bytes to client.
-1. (Connection reset by peer)
Please Help im at a loss to what the problem is!
[global]
getwd cache = yes
socket options = TCP_NODELAY SO_SNDBUF=65536 SO_RCVBUF=65536
IPTOS_LOWDELAY
use sendfile = no
lock spin time = 15
lock spin count = 30
map to guest = bad user
log level = 1
security = user
os level = 64
local master = Yes
time server = Yes
domain master = yes
preferred master = yes
wins support = yes
domain logons = yes
dos filetimes = Yes
workgroup = Family
netbios name = fileserver
server string = Samba Server %v
printcap name = cups
load printers = yes
printing = cups
printer admin = root @adm @is
log file = /var/log/samba/log.%m
max log size = 50
;hosts allow = 192.168.1.0/24 127.0.0.1
interfaces = eth1 lo
bind interfaces only = yes
encrypt passwords = yes
smb passwd file = /etc/samba/smbpasswd
add user script = /usr/sbin/useradd -d /dev/null -g machines -c
'Machine Account' -s /bin/false -M %u
wins proxy = yes
dns proxy = no
logon path =
logon drive = U:
logon script = %U.bat
oplocks = no
level2 oplocks = no
change notify timeout = 300
lpq cache time = 30
winbind uid = 1-2
winbind gid = 1-2
winbind separator = +
oplocks = no
level2 oplocks = no
; deadtime = 60
wins proxy = yes
lpq cache time = 30
change notify timeout = 300
getwd cache = yes
dos filetimes = yes
domain logons = yes
obey pam restrictions = yes
unix password sync = Yes
pam password change = yes
passwd program = /usr/bin/passwd %u
passwd chat = *New*UNIX*password* %n\n *Re*ype*new*UNIX*password* %n\n \
*passwd:*all*authentication*tokens*updated*successfully*
add user script = /usr/sbin/useradd -s /bin/false '%u'
delete user script = /usr/sbin/userdel '%s'
add user to group script = /usr/bin/gpasswd -a '%u' '%g'
delete user from group script = /usr/bin/gpasswd -d '%u' '%g'
set primary group script = /usr/sbin/usermod -g '%g' '%u'
add group script = /usr/sbin/groupadd %g && getent group '%g'|awk -F:
'{print $3}'
delete group script = /usr/sbin/group
