Re: RTM_LOSING: Kernel Suspects Partitioning

Thank you very much for the quick answer. I kept the message quite
short by intention, since I think I already took care of all the
"obvious" stuff - but then again I'm sure I overlooked something.
Especially since I did not install the box myself and the guy who did
is not around.
So, here you go (I changed some information to 'xxx' since the class B
network is connected to the outside world):

The DS10 is connected two times to a layer3 switch (autonegotiation
turned off) - the problem is the same when I connect the client to the
same switch, come from the internet via the gateway (xx.xx.xx.254) to
the fx-uplink on the switch or connect only the two boxes with a hub.
I changed speed and duplex settings on all machines - with no effect,
furthermore the working (win, debian sarge) boxes have perfectly fine
transfer rates to and from the DS10. Then I disabled one interface of
the DS10, since I thought it sends the dropped packets out to the
wrong tu - same. Also with static routing table to the affected
client.


DS10:

# ifconfig -a
lo0: flags=100c89<UP,LOOPBACK,NOARP,MULTICAST,SIMPLEX,NOCHECKSUM>
inet 127.0.0.1 netmask ff000000 ipmtu 4096

sl0: flags=10<POINTOPOINT>

tu0: flags=c63<UP,BROADCAST,NOTRAILERS,RUNNING,MULTICAST,SIMPLEX>
inet xx.xx.xx.5 netmask ffffff00 broadcast xx.xx.xx.255 ipmtu
1500

tu1: flags=c63<UP,BROADCAST,NOTRAILERS,RUNNING,MULTICAST,SIMPLEX>
inet xx.xx.xx.3 netmask ffffff00 broadcast xx.xx.xx.255 ipmtu
1500

tun0: flags=80<NOARP>

tun1: flags=80<NOARP>

#arp -a
Shows the gateway, the connected boxes (NIS, NFS) and when I try to
connect, also my problematic clients show up there correctly.

# netstat -nr
Routing tables
Destination Gateway Flags Refs Use Interface

Route Tree for Protocol Family 26:

Route Tree for Protocol Family 2:
default xx.xx.xx.254 UGS 0 39883 tu0
127.0.0.1 127.0.0.1 UHL 7 1534 lo0
xx.xx/16 xx.xx.xx.254 UGS 2 13453 tu0
xx.xx.xx/24 xx.xx.xx.5 U 5 11370 tu0
xx.xx.xx/24 xx.xx.xx.3 U 2 15261 tu1
xx.xx.xx.3 xx.xx.xx.3 UHL 8 56770 tu1
xx.xx.xx.5 xx.xx.xx.5 UHL 0 0 tu0

This looks somehow strange to me, entry 4 and 5 looks like they could
be causing the problem, but I deleted them inbetween, with no
effect.


# hwmgr -get attr -category network
30:
name = tu0
category = network
sub_category = Ethernet
model = 21143
hardware_rev = 4.1
firmware_rev =
MAC_address = xxx
MTU_size = 1500
media_speed = 100
media_selection = Set by SRM Console
media_type = Unshielded Twisted Pair (UTP)
loopback_mode = 0
promiscuous_mode = 0
full_duplex = 1
multicast_address_list = xxx
interface_number = 1
link =
autoneg_enable = 0
registration_time = Wed Jan 30 21:11:24 2008
user_name = (null) (settable)
location = (null) (settable)
software_module = (null)
state = available
state_previous = unknown
state_change_time = none
event_count = 0
last_event_time = none
access_state = online
access_state_change_time = none
capabilities = 0
indicted = 0
indicted_probability = (null)
indicted_urgency = (null)
disabled = 0
est_seconds = 72472
est_bytesent = 21543732
est_bloksent = 30460
est_mbytesent = 204468
est_mbloksent = 1790
est_deferred = 0
est_single = 0
est_multiple = 0
est_collis = 0
est_unrecog = 0
est_userbuf = 0
est_latecoll = 0
est_excesscoll = 0
est_carrierfail = 0
est_shortcirc = 0
est_opencirc = 0
est_sndlong = 0
est_sendfail = 0
est_bytercvd = 30884980
est_blokrcvd = 113118
est_mbytercvd = 22860197
est_mblokrcvd = 103028
est_overrun = 0
est_sysbuf = 0
est_unaligned = 0
est_longframe = 0
est_shortframe = 0
est_fcsfail = 0
est_badframe = 0
est_symbolerror = 0
est_recvfail = 0
31:
name = tu1
category = network
sub_category = Ethernet
model = 21143
hardware_rev = 4.1
firmware_rev =
MAC_address =xxx
MTU_size = 1500
media_speed = 100
media_selection = Set by SRM Console
media_type = Unshielded Twisted Pair (UTP)
loopback_mode = 0
promiscuous_mode = 0
full_duplex = 1
multicast_address_list = xxx
interface_number = 2
link =
autoneg_enable = 0
registration_time = Wed Jan 30 21:11:24 2008
user_name = (null) (settable)
location = (null) (settable)
software_module = (null)
state = available
state_previous = unknown
state_change_time = none
event_count = 0
last_event_time = none
access_state = online
access_state_change_time = none
capabilities = 0
indicted = 0
indicted_probability = (null)
indicted_urgency = (null)
disabled = 0
est_seconds = 72472
est_bytesent = 3276593
est_bloksent = 8041
est_mbytesent = 223710
est_mbloksent = 2213
est_deferred = 0
est_single = 0
est_multiple = 0
est_collis = 0
est_unrecog = 0
est_userbuf = 0
est_latecoll = 0
est_excesscoll = 0
est_carrierfail = 0
est_shortcirc = 0
est_opencirc = 0
est_sndlong = 0
est_sendfail = 0
est_bytercvd = 35581582
est_blokrcvd = 130582
est_mbytercvd = 22832997
est_mblokrcvd = 102603
est_overrun = 0
est_sysbuf = 0
est_unaligned = 0
est_longframe = 0
est_shortframe = 0
est_fcsfail = 1
est_badframe = 1
est_symbolerror = 0
est_recvfail = 2


Linux client:

$ ifconfig -a
eth0 Link encap:Ethernet HWaddr xxx
inet addr:xx.xx.xx.40 Bcast:xx.xx.xx.255 Mask:
255.255.255.0
inet6 addr: xxx Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:83912 errors:0 dropped:0 overruns:0 frame:0
TX packets:75961 errors:2 dropped:0 overruns:0 carrier:0
collisions:2620 txqueuelen:1000
RX bytes:26074108 (24.8 MB) TX bytes:9998670 (9.5 MB)
Interrupt:17

$ netstat -nr
Kernel IP routing table
Destination Gateway Genmask Flags MSS Window
irtt Iface
xx.xx.xx.0 0.0.0.0 255.255.255.0 U 0 0 0
eth0
0.0.0.0 xx.xx.xx.254 0.0.0.0 UG 0 0 0
eth0

$ arp -a
shows gateway, some boxes from the subnet and the DS10 when I try to
connect.

An ssh connect from my client (.40) to the DS10 (.3) looks like this
(client side):

No. Time Source Destination
Protocol Info
1 0.000000 xx.xx.xx.40 xx.xx.xx.3 TCP
40222 > ssh [SYN] Seq=0 Len=0 MSS=1460 TSV=3311064 TSER=0 WS=7
2 0.000286 xx.xx.xx.3 xx.xx.xx.40 TCP
ssh > 40222 [SYN, ACK] Seq=0 Ack=1 Win=62780 Len=0 MSS=1460 WS=0
3 0.000304 xx.xx.xx.40 xx.xx.xx.3 TCP
40222 > ssh [ACK] Seq=1 Ack=1 Win=5888 Len=0
4 0.013977 xx.xx.xx.3 xx.xx.xx.40 SSHv2
Server Protocol: SSH-2.0-3.2.0 SSH Secure Shell Tru64 UNIX\r
5 0.014035 xx.xx.xx.40 xx.xx.xx.3 TCP
40222 > ssh [ACK] Seq=1 Ack=44 Win=5888 Len=0
6 0.014198 xx.xx.xx.40 xx.xx.xx.3 SSHv2
Client Protocol: SSH-2.0-OpenSSH_4.6p1 Debian-5ubuntu0.1

7 9.551396 xx.xx.xx.40 xx.xx.xx.3 SSHv2
[TCP Retransmission] Client: Unknown (46)[Unreassembled Packet
[incorrect TCP checksum]]
8 9.551603 xx.xx.xx.3 xx.xx.xx.40 TCP
[TCP Previous segment lost] ssh > 40222 [ACK] Seq=532 Ack=41 Win=62780
Len=0
9 9.551627 xx.xx.xx.40 xx.xx.xx.3 SSHv2
Client: Key Exchange Init
10 9.575995 xx.xx.xx.3 xx.xx.xx.40 TCP
ssh > 40222 [ACK] Seq=532 Ack=793 Win=62780 Len=0

The first six packets look fine, and then something happens - but what
and why?

I hope I managed to provide all the needed information this time.

Best Regards & Many Thanks,

Bjoern
.

Relevant Pages

  • XP Pro Client wont accept incoming RDP Connection
    ... Posted the following in our local NG yesterday - despite helpful advice, ... Money-conscious SB Client buys 2 XP Pro clones from local PC shop, ... We reset the local security settings. ... These boxes do run Autocad with a fancy graphics card feeding 28" LCD ...
    (microsoft.public.windows.server.sbs)
  • Re: Security log chaos -
    ... I did not know about the AD client for the 9x boxes. ... do not have it setup on the server at all. ... >win9x machines not getting the correct wins server addresses from the dhcp ...
    (microsoft.public.backoffice.smallbiz)
  • Check box checking & application loads slow
    ... Two problems I'd appreciate any suggestions with [VFP8, XP(dev and client)]: ... At some point when using the check boxes on a pageframe they don't check ... Default data session of the initial formset - dataload method ...
    (microsoft.public.fox.programmer.exchange)
  • RE: Confirming leaving the page.....
    ... one method would be setting a flag at client side when the check boxes are ... This will be a global flag at client side, on page load set it to true ... > javascript window.confirm on unload to stop them. ...
    (microsoft.public.dotnet.framework.aspnet)
  • Re: Locked Form with Synchronized Combo Boxes
    ... Having two combo boxes in and of itself would not usually cause the form to ... But "lock" can mean many different things, ... > I have a form that has a combo box that lets you find a Client record. ... > I added the feature where having picked a client, it loads a second combo ...
    (microsoft.public.access.forms)