Subject : Ethernet Interface FAQ/PSD
Description :
Tips sheet for the Sun Ethernet Interface
1.0: About Ethernet Interface
1.1: Ethernet Introduction
1.2: Pre-Installation notes
1.3 : Built-In Ethernet on Sun
1.4 : Optional Ethernet Interface cards.
1.4.1 : SEC - Sbus Ethernet Controller
1.4.2 : DSBE/S - SBUS FAST DIFFERENTIAL F-W SCSI-2 / BUFFERED ETHERNET Card
1.4.3 : FSBE/S - SBUS FAST SCSI-2/Buffered ETHERNET CARD
1.4.4 : SBE/S - SBUS FAST SCSI-2/Buffered ETHERNET CARD
1.4.5 : SQEC - SBUS QUAD ETHERNET CTRL SBUS CARD
1.4.6 : SUNFASTETHERNET ADAPTER 10/100MBPS SBUS CARD
2.0 Debugging
2.1: Open Boot Prom Diagnostics
2.2: TBD
2.3 etherfind
2.4: nfsstat
2.5: netstat
2.6: ping
2.7: spray
2.8: snoop
2.8: how to turn on debug
3.0 Common How Tos
3.1 Hardware/Software Installation
3.2 : Solaris 2nd Ethenet Sbus Installation.
4.0 Frequently Asked Questions
4.1 Hardware Diagnostics
4.1.1: How to verify that the workstation is recognising harware for the ethernet interfaces
4.1.3: How to set the TPE Link test - for 10baseT configurations.
4.3 : Cabling Information
4.3.1
4.3.2
4.3.3
4.3.4
4.3.6
4.4 Changing primary hostname to 2nd interface.
4.5 I Have mutiple interfaces, but they show up as the same ethernet MAC address
4.6 How to know the actual Ethernet Address?
4.7 03.12Q: Does my Ethernet coax have to be grounded? How?
4.8 Booting over net.
4.9 General Troubleshooting
4.10 : Le and Ie Diagnostic erros and messages:
4.11 How do I define a second TCP/IP address for my ethernet?
5.0: Patches
5.1: Ethernet Driver Patches for SunOS:
5.2: Ethernet Driver Patches for Solaris:
5.3: ARP Patches for SunOS
6.0: Known Bugs & RFEs
6.1: RFEs
6.2: Bugs
7.0: References
7.1 Glossary:
7.2: Important Man Pages
7.3: Sun SRDBs
7.4: Sun Educational Services
7.5: Solaris Documentation
7.6: Third Party Documentation
8.0: Supportability
9.0 Additional Support
1.0: About Ethernet Interface
===========================
이 Tip Sheet document는 SunOS 와 Solaris 에서 구현되는 Ethernet Interface에
관련된 다양한 정보를 묘사하고 있다. 즉 Ethernet Interface에 대한 소개 그리고
가장 일반적인 Configuration 그리고 문제점등에 대해서 다룬다.
이 자료는 Sun 호환 기종에 관련하여 Network 구조의 가장 low Layer, 즉
OSI model에서 Physical, Data Link Layer에 대해서 집중적으로 다루고 있다.
1.1: Ethernet Introduction
===========================
Network는 최소한 하나 이상의 W/S의 집합이며 같은 Type 또는 다른 Type 간의
서로 물리적인 정렬에 의해 연결되어 있다.
Ethernet은 LAN 환경에서 w/s을 연결하는 기술의 하나이다. Ethernet/802.3은
현재 multi-mode fiber(10baseFL) 뿐만 아니라 thin-net coaxial cable (10base2),
thick-net coaxial cable (10base5), UTP-unshielded twisted-pair STP-Sheilded
twisted pair (10baseT) 에서도 Running 된다.
Ethernet에서 최대 Data 전송속도는 10 megabits/second(raw data)이다.
Physical layer에서 이러한 구현은 IEEE 802.3 과 Ethernet 사양에 적용되는
10 Mbps CSMA/CD protocol-based network interface을 위해 Media Access Control
Control (MAC) 기능을 제공한다.
Ethernet은 transceivers(MAU - Medium Attachment Unit), transceiver cables(AUI),
Shielded coaxial cables 그리고 일반적으로 사용되는 Unshieled Twisted Pair(UTP)
그리고 Shielded Twited Pair(STP)등을 사용하여 상호 W/S 간에 통신한다.
대부분의 workstation들은 transceiver cables(AUI) 또는 Twisted Pair(TPE)등으로
network에 바로 접속되어 있다.
1.2: Pre-Installation notes
=============================
만약 Cabling작업시 10baseT TPE Connect Type 이라면 전화 Line Connect Type과
비슷하기 때문에 주의를 요한다. 잘못 Cabling이 되면 computer 뿐만 아니라
network에 손상을 입힐수 있다.
그리고 10base2 또는 10base5에서 AUI transciever를 사용할 때 외장 transceiver
(MAU)와 연결하기 전에 system power를 끄기를 권장한다. 초기 MAU Power가
system에 손상을 줄수 있다.
1.3 : Built-In Ethernet on Sun
===========================
Sun4c Architecture:
Sparc1, Sparc1+, Sparc2, Sparc IPC, Sparc IPX, Sun4/300 :
Hardware provides ethernet capabilities with a built-in 15-pin DB-15 AUI Connection.
Sun4m Architecture:
SPARCstation-5, SPARCstation-10, SPARCstation-20 and,SPARCclassic systems:
Comes with two types of Ethernet interfaces.
The Symbol - -is used to identify the ethernet ports.
RJ45 connector- for 10BaseT (Twisted- Pair Ethernet or TPE)
AUI connector- for 10Base2 (Thick Ethernet) 10base5(Thin Ethernet)
transceiver dependent.
The AUI port uses a Mini-26 connector which is below the Parallel port
and requires an AUI adaptor cable -X981A (530-2021). SS10 & LX splits
audio and AUI on a mini-26/db15 AUI adaptor X987A (530-1856).
하나의 Ethernet Controller에서 두개의 Type AUI, TPE port가 있지만
단지 하나만이 연결된다. 먼저 AUI port가 test 되고 다음 AUI test 가
실패하면 TPE port를 test하고 active 시킨다.
The SPARCstation Voyager:
includes does not support AUI Ethernet connections.
Sun4d Architecture:
For SparcServers :
The main network interface is an SBus card on the system master board,
which is the system board in card cage slot 0.
Solaris X86 Architecture:
Dependent on Network Interface card (NIC) installed.
See X86 documentation on supported NIC's.
Transceiver will vary with the type of network.
1.4 : Optional Ethernet Interface cards.
=======================================
지원이 되는 SBus LANCE Ethernet controller는 여러가지 다양하다.
Buffer가 있는 ethernet card는 Ethernet DMA transfer buffering을 위해
128-KB의 memory를 포함하고 있다. SunOS 와 Solaris Software는
"style 2" Data Link Service를 제공하는 Ethernet device driver를 포함한다.
Interface cards utilizing IEEE 802.3 CSMA/CD
Second Ethernet (all SBus and only 3/E VME)
SBus SCSI/Buffered Ethernet Card (SBE/S)
SBus Fast SCSI Buffered Ethernet Card (FSBE/S)
SBus Differential SCSI Buffered Ethernet Card (DSBE/S)
SBus Quad Ethernet Controller (SQEC)
SunFastEtherNetSBus(TM)
Other SunLink Options:
Contact your Sun Representive, SunExpress or Other PSD's for infomation on:
SBus Fiber Distributed Data Interface (FDDI/S)
SBus High-Speed Interface (HSI/S)
SBus Token Ring Interface (TRI/S)
Integrated Services Digital Network (ISDN)
이하에서 List되는 Card는 모든 Sun 호환 workstations 그리고 SBus slot을 갖는
server에서 지원된다. 그리고 Solaris 1.1.1 operating system revision 또는
그 이상에선 OpenBoot PROM Revision이 2.0 또는 그 이상이어야 한다.
1.4.1 : SEC - Sbus Ethernet Controller
==============================
* X453 & X450 SECOND ETHERNET CONTROLLER
X540 PN 501-1450
X543 PN 501-1881
Ethernet Connectors:
BNC
AUI
Switches -
BNC - cheapernet 1-6 ON, 7-8 OFF.
AUI - ethernet : 1-6 OFF, 7-8 ON (+12vdc)
Notes:
Normally, the +12VDC is turned off. If you need to turn the +12VDC on,
you can do so if there is only one internal hard disk drive installed
in a two hard disk drive system. For SPARCstation IPX systems with the
internal 424 Mbyte hard disk drive installed, the +12VDC on option
is not allowed.
Not Supported in SPARCstation 5, 10, 20 Systems
1.4.2 : DSBE/S - SBUS FAST DIFFERENTIAL F-W SCSI-2 / BUFFERED ETHERNET Card
=================================
X1052 SBUS FAST DIFFERENTIAL Fast-Wide SCSI-2 / BUFFERED ETHERNET CARD (DSBE/S)
PN 501-1902
Ethernet connector:
- RJ45 10baseT
Jumpers:
J0302 = OUT : enable TPE link Test
J0302 = IN : disable TPE link Test
1.4.3 : FSBE/S - SBUS FAST SCSI-2/Buffered ETHERNET CARD
=======================================================
X1053 SBUS FAST SCSI-2/Buffered ETHERNET CARD (FSBE/S)
PN 501-2015
Ethernet connector- RJ45 10baseT
Jumpers:
J0302 = OUT : enable TPE link Test
J0302 = IN : disable TPE link Test
1.4.4 : SBE/S - SBUS FAST SCSI-2/Buffered ETHERNET CARD
========================
X1054 SBUS FAST SCSI-2/Buffered ETHERNET CARD (SBE/S)
PN: 501-1869
Ethernet connector:
14-pin champ connector
10base5 adaptor 530-1812
10baseT Adapter 530-1813
Jumpers:
J0202 In Enable 10base5 Ethernet
J0302 in Enable 10base-T Ethernet (tpe false)
Notes:
SPARCclassic 계열과 SPARCstation LX 계열에서는 지원되지 않는다.
SBE/S card 상에서 Twisted Pair Ethernet interface는 system 과 Ethernet Hub
간 cable connection을 검사하는 Link Integrity Test를 지원하지 않는다.
만약 SBE/S card 상에서 Link Integrity Test 가 10BASE-T Hub에서 enable
되어 있다면 network가 기능하지 않을 것이다. no carrier와 같은 Failure,
rlogin 또는 ping command에 반응하지 않을 것이다. 10BASE-T Hub 상에서는
Link Integrity Test를 Disable 하라.
If the Link Integrity Test cannot be disabled on the Hub, a workaround
is to use Adapter Cable 530-1812-01 and an AUI to 10BASE-T Adapter, such
as the AT&T StarLAN 10 Network AUI Adapter.
1.4.5 : SQEC - SBUS QUAD ETHERNET CTRL SBUS CARD
=================================================
* X1058 SBUS QUAD ETHERNET CTRL
PN 501-2655
SBus Quad Ethernet Controller (SQEC)는 4개의 fully-buffered
Ethernet port를 제공한다. Ethernet channel은 각각 link integrity LED
를 포함하면서 독립적으로 10 Mbps로 작동된다.
Ethernet connectors:
RJ45 10baseT
Notes:
TPE Link Integrety Test는 Open Boot Prom을 통해 enable/disable된다.
1.4.6 : SUNFASTETHERNET ADAPTER 10/100MBPS SBUS CARD
====================================================
* X1056 SUNFASTETHERNET ADAPTER 10/100MBPS SBUS CARD
100BaseT Lan Technology. SunFastEthernet(TM) adapter cards는 초당
10/100 Mbit에서 작동하는 dual-mode high-speed Ethernet SBus card이다.
만약 OBP version이 2.3 이전이면 OBP version을 upgrade하여야 한다.
10 Mbps 또는 100 Mbps speed로 auto-sense하는 특성을 갖는다.
Fast Ethernet Alliance 사양에 부응하는 third party Fast Ethernet
hub/switch를 필요로 한다.
Connectors:
RJ45 :
MII : Media Independent Interface (MII), 다양한 type의 wiring에
걸쳐 작동이 되도록 external third party transceiver에 연결되는
Interface. TPE Link Integrety Test는 Open Boot Prom을 통해
enable/disable된다.
Notes:
SunOS 4.1.3 or Solaris 2.3 (H/W 5/95) or higher is required.
SunOS requires SunFastEthernet 1.1 Software and will only work on
sun4m platforms.
Solaris requires the SUNWbmac package.
2.0 Debugging
==========================
2.1: Open Boot Prom Diagnostics
=================================
Ethernet 과 SCSI interfaces, system memory, diskette drive controller등의
test를 위해 Open Boot Prom FORTH Diagnostic을 사용할 수 있다.
ok prompt ( FORTH Monitor)에서 FORTH Diagnostic을 수행해야 한다.
FORTH Diagnosticd은 boot PROM에 실장되어 있다.
test net, test net-aui, test net-tpe
watch-net, watch-aui, watch-tpe, and watch-net-all
Use watch-tpe to monitor the 10BaseT (TPE) connection or watch-aui
to monitor the 10Base5 (thick Ethernet) connection.
example:
ok watch-net-all
/iommu@f,e0000000/sbus@f,e0001000/lebuffer@1,40000/le@1,60000
Internal loopback test -- succeeded.
External loopback test -- succeeded.
Looking for Ethernet packets.
'.' is a good packet. 'X' is a bad packet.
Type any key to stop.
...............................................................
....................(keyboard key pressed here)
/iommu@f,e0000000/sbus@f,e0001000/ledma@f,400010/le@f,c00000
Internal loopback test -- succeeded.
External loopback test -- succeeded.
Looking for Ethernet packets.
'.' is a good packet. 'X' is a bad packet.
Type any key to stop.
...................................................
Ethernet Transceiver MAU의 Power와 Data, LINK LED가 ON 되어 있는지를
검사하라. 대부분의 10baseT HUB는 또한 link light를 갖고 있다.
test net-tpe Performs an internal and external loopback test
on the TPE interface.
A cable must be connected to the system TPE port and to a TPE hub or
the test will fail the external loopback phase.
If the tpe-link- (disabled), the external loopback test will appear
to pass even if a cable is not connected.
On sun4 use the extended diagnostics can be used in OBP.
Enter "x" at the monitor prompt.
2.2
2.3 etherfind
=================
etherfind program은 Ethernet상의 packet을 찾고 주어진 expression에 맞게
정보를 print한다. -v option은 network problem을 분석할 수 있다.
etherfind를 사용하므로써 database server와 client간에 어떻게 interact하는
지를 나타낼수 있다.
2.4 nfsstat
=================
이 utility는 NFS retransmission 비율을 측량할 수 있다. 다양한 option은
problem의 범위를 구분할 수 있다. nfsstat -z는 특정 시간동안 모아진
통계를 볼 수 있다.
nfsstat -rc on the client as shown here.
# nfsstat -rc
Client rpc:
calls badcalls retrans badxids timeouts waits newcreds
63472 0 11 0 11 0 0
badverfs timers toobig nomem cantsend bufulocks
0 39 0 0 0 0
time-out은 RPC time out 건수를 알린다. client는 재전송을 특정횟수 시도한후
에도 reply가 없으면 빠진다. client가 server로 부터 RPC time-out 시간내에
reply를 받지 못한다는 것은 다음과 같음을 의미할 수 있다.
* Packets are lost on the network.
* Network congestion slows packet transmission.
* Packets are dropped at server's network interface or in protocol
queues.
* Server takes too long to handle the request.
The time-out values should be small in relation to the total number of calls
made. The request xid is retained across transmissions. If timeout and
badxid are almost the same number, it implies that most retransmitted
requests are being successfully processed.
badxid - Duplicate responses from server, the number of retransmitted
requests successfully processed. If badxid = 0, it means that retransmitted
requests are being dropped.
retrans - Retransmission count. This number reports the number of times an
RPC is retransmitted at the request of the caller. Retransmission rates that
exceed 5-10% of the total number of clients NFS calls indicate a problem
that needs further investigation. The retransmission count is affected by
NFS dynamic retransmission algorithm. Dynamic retransmission is enabled when
the client and server are on different networks. Since bridges do not
trigger dynamic retransmission, you must anticipate bottlenecks at routers
and gateways.
badcalls - Time-outs resulting in RPC error.
Determine when excessive symbolic links are increasing the number of RPC
calls with nfsstat -ns as shown next.
readlink - Shows the number of reads to symbolic links. If the percentage of
RPC calls approaches 10%, this could mean there are too many symbolic links.
See Chapter 4, "Tips on Fine Tuning" for techniques to reduce symbolic
links.
For an example of network debugging see Managing NIS and NFS. [19])
2.5 netstat
=======================
This tool shows the status of the network by the number of network
collisions, by identifying the active sockets, routing, and traffic. You can
see cumulative statistics for packets transferred, errors, collisions,
network addresses for the interface, and the maximum transmission unit
(mtu), depending on the option selected.
The next example shows netstat.
which gives cumulative statistics since the last boot.
# netstat -i
Name Mtu Net/Dest Address Ipkts Ierrs Opkts Oerrs Collis Queue
lo0 8232 127.0.0.0 localhost 11174 0 11174 0 0 0
le0 1500 129.151.21.0 hodware 302265 5 119032 1 9406 0
The first column shows the number of packets handled, followed by the errors
made in transmitting or receiving them. Statistics are shown for selected
network interfaces, in this example e0, and show the total number handled
by all interfaces.
colls - Reports the number of collisions since the network booted. If output
colls is greater than 5% of output packets on one system, check the network
wiring to that system. When output colls is consistently greater than 5% of
output packets on most of the systems, it is time to consider subdividing
the network into separate networks connected by a router or a bridge. A
router can be a mixed function or communications server, or a workstation
with two network interfaces can be dedicated to providing router
functionality.
A high collision/defer rate is an indication of an overloaded Ethernet.
After checking for problems with hosts on the Ethernet, the network should
be partitioned if the network-wide collision rate routinely exceeds 10%. If
the average usage level exceeds 30% it indicates an overloaded Ethernet.
Find out what is generating the traffic and change it. Check that diskless
workstations have enough memory.
Example: netstat -k !more -then search for interface
le0 qeX= qe1 ro qe2, etc.
and the first section you see whould be
similiar to:
...skipping
le0:
ipackets 757651 ierrors 1 opackets 186505 oerrors 25 collisions 28700
defer 6678 framming 1 crc 1 oflo 0 uflo 0 missed 0 late_collisions 1
retry_error 23 nocarrier 0 inits 13 nocanput 625 allocbfail 0
notes form na.hostperf
The way collision count is tallied depends on the Ethernet
chip. The Intel (ie) chip counts every occurrence, while
the Lance (le) chip counts collision on a per-packet basis.
With the Lance chip, if a packet is successfully sent after
the first retry, the collision count is one. If a packet is
successfully sent after two or more retries, the chip
returns a maximum count of two collisions. For example, if
packet a had four retries and sometime later packet b had
five retries, the cumulative collision count is four - two
for each packet - not the real count of nine.
2.6 ping
===============================
The command ping will tell you if a machine is active and reachable over
your network. Use ping -s to show the transit times (round trip times) and
packet losses. When using ping for fault isolation, test the local host with
ping to make sure the local network interface is running. One test is to use
the ping command These 8000 packets get split into 1500-byte Ethernet packets,
thus simulating the kind of load NFS places on the network. The large packet
size is important, as some kinds of failures seem not to show up with smaller
packets. See the man page ping (8C). An example of using ping to isolate
network problems can be found in Managing NFS and NIS. [20])
2.7 spray
==================================
reliability of your specified
packet sizes, or to determine the largest reliable size of packet for your
network. it works over UDP/IP (nfs uses udp)
and does not
use any flow control. spray should be able to tell you whether packets are
getting dropped or delayed. Try the following command:
spray -d 750 -l 1514 -c 1000 nobs
sending 1000 packets of length 1514 to nobs ...
no packets dropped by nobs
803 packets/sec, 1216802 bytes/sec
This sends 1000 packets of 1514 byte size with a 750 millisecond delay.
between the bursts.
2.8 Snoop
===============================
It turns a Sun machine into a basic network sniffer,
able to detect Ethernet network problems. Snoop also captures a certain
number of Ethernet packets and allows you to trace the calls from each
client to each server, as shown in the next figure.
Snoop captures packets from the network and displays their contents. It uses
both the packet filter and buffer modules of the NIT interface to provide
efficient capture of packets from the networks. Captured packets can be
displayed as they are received, or can be saved to a file for later
inspection.
Snoop can display packets in a single-line summary or in expanded form. In
summary form, only the data pertaining to the highest level protocol is
displayed. For example, an NFS packet will have only NFS information
displayed. The underlying RPC, UDP, IP and Ethernet frame information is
suppressed, but can be displayed if either of the verbose options are
chosen.
Snoop is a useful tool if you are considering subnetting, since it provides
load statistics based on each server. The program is capable of breaking the
packet header out in order to debug it, and to see what might be the source
of incompatibility difficulties.The next screen sample shows typical DLC and
IP headers as captured in Snoop.
2.8 : How to turn on debug
===================================
This will send verbose messages from the le driver to
/var/adm/messages, two lines for each message.
There are two methods for turning on ledebug in Solaris 2.x
1) Add the following entry to /etc/system and reboot:
set le:ledebug=1
2) Issue the following commands on as root:
# adb -kw /dev/ksyms /dev/mem
ledebug/X
(will display current)
ledebug/W 1
(will set current to 1 - be sure to use W not w)
On 4.x, adb:
Issue the following commands on as root:
# adb -kw /vmunix /dev/mem
ledebug/W 1 (to set the in memory copy)
ledebug?W 1 (to set the on disk copy)
3.0 Common How Tos
3.1 Hardware/Software Installation
3.2 : Solaris 2nd Ethenet Sbus Installation.
1. Install a Sun Ethernet S SBus card into the system as described in
the instructions and manual provided with the card. Also refer to
your Installation Guide for your model of Workstation
2. Close the system and connect the cables. Double check connection to
AUI port or twisted-pair Ethernet port).
3. Boot the system and become superuser.
4. Touch /reconfigure prior to reboot.
5. Shut down system, power off, install board (S-bus) using
static precautions (a disposible Wrist strap should be included
in kit)
6. boot -r from the OK prompt if Step #1 not done
7. look for "le" interface number of board (look thru dmesg output).
or make an educated guess, and use it in steps #5, #6 and #7.
8. /sbin/ifconfig le1 plumb (or le2, or le3, etc.)
IMPORTANT!! This step tells Streams that card exists.
9. Try ifconfig -a - the card SHOULD appear.
If not already so done,
create /etc/hostname.le1 (or le2, or le3, etc.).
10. You can try to bring up the card manually (with ifconfig le1
inet ...),
or reboot the system at this point.
11. After rebooting system or bringing up the card manually, try
pinging hosts on the network attached to the add-on ethernet's
network. The add-oo ethernet interface should respond correctly,
should display in response to an ifconfig -a. The le device
should display in the output from prtconf, and should display
in the output from dmesg.
4.0 Frequently Asked Questions
4.1 Hardware Diagnostics
====================================
4.1.1: How to verify that the workstation is recognising harware for the ethernet interfaces.
==================
On Sparc Systems you can check the devices at the OBP ok prompt.
devalias - displays current device aliases.
show-sbus - the card will be identified in a Slot #.
show-devs - shows the device tree
/iommu@0,10000000/sbus@0,10001000/le@0,c00000
/iommu@0,10000000/sbus@0,10001000/ledma@4,8400010
/iommu@0,10000000/sbus@0,10001000/ledma@4,8400010/le@4,8c00000
On boot-up you will notice Entries like:
Solaris 2.x -
le0 is /iommu@0,10000000/sbus@0,10001000/le@0,c00000
ledma0 at sbus0: SBus slot 4 0x8400010
le1 at ledma0: SBus slot 4 0x8c00000 sparc ipl 6
le1 is /iommu@0,10000000/sbus@0,10001000/ledma@4,8400010/le@4,8c00000
lebuffer0 at sbus0: SBus slot f 0x40000
le0 at lebuffer0: SBus slot f 0x60000 sparc ipl 6
le0 is /iommu@f,e0000000/sbus@f,e0001000/lebuffer@f,40000/le@f,60000
lebuffer1 at sbus0: SBus slot 0 0x40000
le1 at lebuffer1: SBus slot 0 0x60000 SBus level 4 sparc ipl 7
le1 is /iommu@f,e0000000/sbus@f,e0001000/lebuffer@0,40000/le@0,60000
SunOS 4.x -
le0 at SBus slot 0 0xc00000 pri 5
If it is not seen with SunOS4.x, it is possible that the kernel may need reconfigured.
check the /usr/sys/'arch'/conf/YOURKERNEL for lines that may be commented out.
Examples (depending on System architecture):
device-driver le
device ie1 at vme24d16 ? csr 0xe88000 priority 3 vector ieintr 0x75
device le0 at obio ? csr 0xf9000000 priority 3
device ie0 at obio ? csr 0x6000000 priority 3
The command prtconf will display the device tree on your system
# prtconf -p
Node 'iommu'
Node 'sbus'
Node 'espdma'
Node 'ledma'
Node 'le'
Node 'le'
4.1.3: How to set the TPE Link test - for 10baseT configurations.
=============================================
The Link Integrity Test tests communications between the ethernet port in the
server and the nearest network hub. If the test is set differently at the card
and the hub, communication is not possible. (A monitor attached to the server will
display the message "no carrier" or "not responding" or similar errors.
Most modern 10baseT hubs support TPE Link Integrety Test, It may be advisable to
upgrade a Hub if it does not confirm to specifications.
If the tpe-link- test? parameter is false (disabled), the external loopback test
will appear to pass even if a cable is not connected.
TPE Link test can be Disabled(In)or Enabled(Out) with the J0302 Jumper on most
network SBUS cards,
It is preferable to disable (or enable) the test through software, as shown below.
For a single network, disable the test with one of the following commands.
At the OpenBoot prompt, enter: set-tpe-test 0
At the ok prompt, enter: false set-tpe-test
If it becomes necessary to re-enable testing at the server, use one of the following commands.
At the OpenBoot prompt: set-tpe-test 1
At the ok prompt: true set-tpe-test
For a system connected to more than one network, it is recommended that you program the system
NVRAM with statements of the type:
This is only required on special cases and older Hubs.
-------------------------------------------------------------------------------
example: "Ethernet-node1-pathname" would be the physical device
false " 3
2 <--[TX-]--------[RX-]--> 6
3 <--[RX+]--------[TX+]--> 1
6 <--[RX-]--------[TX-]--> 2
4.4 Changing primary hostname to 2nd interface.
For newly installed systems or if a system is considered re-configurable,
see the "sys-unconfig" command, this will allow the root user to reassign
the host information upon reboot.
There is also the ability to identify a system by using alias names in the hosts file.
Each name must be unique. Having duplicate names for different IP numbers will result in the
system locating the first name in the list.
example: /etc/hosts
200.100.20.10 sparc5 server1 mailhost
In Solaris /etc/hosts is a link to /etc/inet/hosts, the order of hosts lookups are identified in
/etc/nsswitch.conf
hosts: files nis dns
Solaris:
the hostname of the system is identified in the boot scripts
hostname="`shcat /etc/nodename 2>/dev/null`"
The /etc/nodename should be the same name as your primary interface /etc/hostname.XX#
Change /etc/nodename to the same as /etc/hostname.le1
If no longer using the oldsystem name -
verify that the files in /etc/net/ticlts, ticots, ticotsord
reflect new nodename
Use the grep command to check for any reference to the old system name.
# grep oldsystemname /etc/auto*
# grep oldsystemname /etc/host*
# grep oldsystemname /etc/vfstab
# grep oldsystemname /etc/dfs/dfstab
With NIS check:
# ypcat hosts
With NIS+ check:
# niscat hosts.org_dir
SunOS
1. Boot the system and become superuser.
2. Edit the /etc/rc.boot file: (example - using le1 interface as primary hostname)
a. Find the line in the /etc/rc.boot file with this entry:
hostname="`shcat /etc/hostname.??0 2/dev/null`"
b. Change the hostname.??0 to hostname.??1 and save the changes.
The entry should now appear as:
hostname="`shcat /etc/hostname.??1 2/dev/null`"
3. Change the name in the /etc/hostname.le1 file to the new name.
verify the that the hostname in /etc/hostname.le1 is the name on a line
by itself and correct naming convention.
If disabling the le0, you can:
Type mv /etc/hostname.le0 /etc/hostname.le1 and press Return.
4. Check local /etc/hosts file. At minimun Your hostname line should look something like:
127.0.0.1 localhost
200.100.20.10 myprimaryhostname loghost
4. Halt the system and reboot.
The system will now use the SBus card's Ethernet interface.
4.5 I Have mutiple interfaces, but they show up as the same
ethernet MAC address as my built-in interface. Setting NVRAM
parameter local-mac-address?=true does not seem to effect the address.
The ethernet address is a station address, not an interface address.
All interfaces of one "station" (e.g. a workstation) share the
same ethernet address. Multiple interfaces do
not get different ethernet addresses. The drivers in Sun systems get
the MAC address from the PROM on the system. The MAC address does not
come from the ethernet chip or interface hardware. There is just one
ethernet MAC address for all interfaces on a system.
The NVRAM parameter local-mac-address? is presently reserved.
Some network configurations require Unique MAC addresses.
To change the Ethernet address for interface use ifconfig(1M).
example% ifconfig le0 ether aa:1:2:3:4:5
This change can be set permenantly in /etc/rc2.d/S72inetsvc
using .enet-addr Displays the current Ethernet address
4.6 How to know the actual Ethernet Address?
use "banner "command at Open Boot Pr ok prompt.
or at OK prmpt use:
".enet-addr " Displays the current Ethernet address
Use the command dmesg and or examine file /var/adm/messages
Example output: Ethernet address = 8:0:20:73:f:cc
ifconfig returns the ether address, only if run by root.
Example output from ordinary user:
% ifconfig -a
le0: flags=63
inet 129.155.37.44 netmask ffffff00 broadcast 129.155.37.0
Example output from superuser (root):
# ifconfig -a
le0: flags=63
inet 129.155.37.44 netmask ffffff00 broadcast 129.155.37.0
ether 8:0:20:9:ae:a2
4.7
03.12Q: Does my Ethernet coax have to be grounded? How?
A: The 10Base2 spec says the coax MAY be grounded at one
and only one point, while the 10Base5 spec says the coax SHALL be
grounded at one and only one point.
Grounding your coax is generally you should absolutely
install cabling according to your electrical codes.
4.8
Booting over net.
If the auto-boot? parameter is set to true (default), and the diag-switch? parameter is set to true (not the default), the operating system is booted using the device alias net.
Boot operating system (/kernel/unix) from network automatically only if diag-device
is set to net (default setting)
4.9
General Troubleshooting:
---------------------
* Network hardware problems, such as improper network termination, a
misseated board, a loose connector, or degenerated media causing
network noise.
* A poorly configured operating system kernel that occupies more than
the necessary amount of memory
* The mounted file system - the way utilities are distributed can cause
successive RPC (Remote Procedure Calls) calls
* Client-side bottlenecks leading to excessive paging over the network
* Server CPU or disk I/O bottlenecks resulting in slow access times
* NFS request retransmissions creating network congestion
4.10 : Le and Ie Diagnostic erros and messages:
=====================================================
PECIFIC):
le%d: transmitter frozen -- resetting
A bug in the LANCE
chip has caused the chip's transmitter section to stop. The
driver has detected this condition and reinitialized the
chip.
le%d: out of mbufs: output packet dropped
The driver has
run out of memory to use to buffer packets on output. The
packet being transmitted at the time of occurrence is lost.
This error is usually symptomatic of trouble elsewhere in
the kernel.
le%d: stray transmitter interrupt
The LANCE chip has signalled that it
completed transmitting a packet but the
driver has sent no such packet.
le%d: LANCE Rev C/D Extra Byte(s) bug; Packet dropped
The LANCE chip's internal silo pointers have become misaligned.
This error arises from a chip bug.
le%d: trailer error
An incoming packet claimed to have a
trailing header but did not.
le%d: runt packet
An incoming packet's size was below the
Ethernet minimum transmission size.
le%d: Receive buffer error - BUFF bit set in rmd
This error ``should never happen,''
as it occurs only in conjunction
with a LANCE feature that the driver does not use.
le%d: Received packet with STP bit in rmd cleared
This error ``should never happen,''
as it occurs only in conjunc-
tion with a LANCE feature that the driver does not use.
le%d: Received packet with ENP bit in rmd cleared
This error ``should never happen,''
as it occurs only in conjunc-
tion with a LANCE feature that the driver does not use.
le%d: Transmit buffer error - BUFF bit set in tmd
Excessive bus contention has prevented the LANCE chip from gathering
packet contents quickly enough to sustain the packet's
transmission over the Ethernet. The affected packet is
lost.
le%d: Transmit late collision - Net problem?
A packet collision has occurred after the channel's slot
time has elapsed. This error usually indicates faulty hardware else-
where on the net.
le%d: No carrier - transceiver cable problem?
The LANCE chip has lost input to its carrier detect pin
while trying to transmit a packet.
le%d: Transmit retried more than 16 times - net jammed
Network activity has become so intense that sixteen successive
transmission attempts failed, causing the LANCE chip to give
up on the current packet.
le%d: missed packet
The driver has dropped an incoming
packet because it had no buffer space for it.
le%d: Babble error - sent a packet longer than the maximum length
While transmitting a packet, the LANCE chip has
noticed that the packet's length exceeds the maximum allowed
for Ethernet. This error indicates a kernel bug.
le%d: Memory Error!
Ethernet chip memory access timed out
The LANCE chip timed out while trying to acquire the bus for
a DVMA transfer.
le%d: Reception stopped
Because of some other error, the
receive section of the LANCE chip shut down and had to be
restarted.
le%d: Transmission stopped
Because of some other error, the
transmit section of the LANCE chip shut down and had to be
restarted.
le0: no upper interface for 802.3 packet
This is probably due to the past install an old release
(4.0 or earlier) of SunLink OSI or DNI. This is caused
by something on your net broadcasting an 802.3 packet. Note that
0060 and 0028 are not legal Ethernet packet types. This implies that
they are 802.3 packets. The standard Ethernet driver thows these away;
the one delivered with OSI and DNI contains code for handling 802.3
packets. The message is generated when no upper layer client for this
packet type is known.
[OTHER MISC/GENERAL ERRORS]
ie0 spurious interrupt
ie0 command not accepted
panic: iechkcca
(see June STP, section #2)
Iechkcca panic associated with hardware addressing and
possible priority problems/conflicts. Review the following:
1) Hardware problem
2) identify cause (offending hardware)
3) If Sun hardware, then
-check board jumpers, switches, etc
-swap board out
4) If non-Sun board
-we don't support
-mention consulting help available
ie0: WARNING if_snd full
This message probably shouldn't be passed on to the customer,
because all it means is that the driver wants to send packets
and the output queue filled up before all of them got out. Some
of these type messages (like WARNING: ie0: requeueing) are being
removed from newer releases because, although they indicate a high
traffic situation, they are not necessarily fatal. This queue
overflow can occur because of a hardware failure on the cpu or a
coax or tranceiver problem, but if it's only happening once in awhile
it probably is just indicating a busy net.
nd0: output error 55
(related to above error ie0: WARNING if_snd full)
comes from the next layer up in software and is a direct result of
the first error. You were right about the /usr/include/sys/errno.h
description of "no buffer space available". The driver got backed up
when the output queue filled up, but the kernel was still allocating
buffer space for more packets and just ran out of mbufs, resulting in
this error.
le0: Memory error!
==================
This message indicates that the network interface encountered an
access time-out from the CPU's main memory. There is probably
nothing wrong except system overload.
If the system is busy with other processes, this error can occur
frequently. If possible, try to reduce the system load by
quitting applications or killing some processes.
The Lance Ethernet chip timed out while trying to acquire the bus
for a DVMA transfer. Most network applications wait for a
transfer to occur, so generally no data gets lost. However, data
transfer might fail after too many time-outs.
For more information about the Lance Ethernet chip, see the
le(7D) man page.
le0: No carrier-- cable disconnected or hub link test disabled?
===============================================================
Standalone machines with no Ethernet port connection get this
error when the system triesto access the network. If the
Ethernet cable is disconnected, SPARC machines with the sun4m
architecture usually display this message, whereas machines with
the sun4c architecture usually display the "le0: No carrier--
transceiver cable problem" message instead. If the Ethernet cable
is connected, this message could result from a mismatch between
the machine's NVRAM settings and the Ethernet hub settings.
If this message is continuous, try to save any workto local
disk.
When a machine is configured as a networked system, it must be
plugged into the Ethernet with a twisted pair J45 connector.
If the Ethernet cable is plugged in, find out whether or not the
Ethernet hub does a Link Integrity Test. Then become superuser to
check and possibly set the machine's NVRAM. If the hub's Link
Integrity Test is disabled, set this variable to false.
# eeprom | grep tpe tpe-link-test?=true # eeprom 'tpe-link-
test?=false'
The default setting is true. If for some reason tpe-link-test?
was set to false,and the hub's Link Integrity Test is enabled,
set this variable to true.
le0: No carrier-- transceiver cable problem?
============================================
Standalone machines with no Ethernet port connection If
the network test fails, check the Ethernet port, card, fuse, and
cable, replacing them if necessary. Also check the twisted pair
port to make sure it is patched to the correct subnet.
For more information on packets, see SPARC: Installing Solaris
Software. If you are using the AnswerBook, "ARP/RARP" isa good
search string.
4.11 How do I define a second TCP/IP address for my ethernet?
You can define up to 255 "logical addresses" for a given
network interface on Solaris. There is no supported method
to do so on SunOS 4.1.X.
Here is an example of adding an le0:1 network interface:
# ifconfig le0:1 10.0.0.1 netmask 255.255.255.0 broadcast 10.0.0.255
^^
# ifconfig -a
lo0: flags=849 mtu 8232
inet 127.0.0.1 netmask ff000000
le0: flags=863 mtu 1500
inet 129.151.21.82 netmask ffffff00 broadcast 129.151.21.255
ether 8:0:20:23:61:d9
le0:1: flags=842 mtu 1500
inet 10.0.0.1 netmask ffffff00 broadcast 10.0.0.255
netstat -i
Name Mtu Net/Dest Address Ipkts Ierrs Opkts Oerrs Collis Queue
lo0 8232 127.0.0.0 localhost 235599 0 235599 0 0 0 /net/osseast.east/sunsolvehome/data/patches/SunOS: No such file or directory
le0 1500 129.151.21.0 jedi 2800863 7094 707901 17 92577 0
le0:1 1500 10.0.0.0 10.0.0.1 0 0 0 0 0 0
5.0: Patches
============
The following is the list of all of the Ethernet related patches for
4.1.3, 4.1.3_u1, 4.1.4, 5.3 and 5.4. If you are having Ethernet
problems, installing the patches is a good place to start, especially
if you recognize the general symptoms noted below.
In order for a machine to be stable, all of the recommended patches
should be installed as well. The list of recommended patches for your
operating system is available from sunsolve1.sun.com.
5.1: Ethernet Driver Patches for SunOS:
==================================
SunOS 4.1.3
100768-02 (le)
SunOS 4.1.3_U1
101954-06 (le)
102143-01 (ie)
SunOS 4.1.4
102430-02 (le)
5.2: Ethernet Driver Patches for Solaris:
==================================
Solaris 2.3
101306-10 (le)
101820-07 (be)
102272-01 (sun4m-iommu)
101306-10 (qe)
Solaris 2.4
102001-08 (le, be, and qe SuperJumbo Patch)
102038-02 (Sparc5-iommu)
102332-01 (sun4m-ledma)
)
102729-01 (point patch for SS5 & SS20)
Solaris 2.4x86
atch-ID# (le)
102065-02 (smc)
Solaris 2.5
There is a "T" patch, T102979-01, for the fastethernet driver for 2.5.
It fixes one bug: 1225810 memory leakage in be driver
5.3: ARP Patches for SunOS
--------------------------
4.1.1 4.1.2 4.1.3:
server not responding due to limits of arp
Corrects problems caused by the arp table being limited to 512 entries.
Should be installed if you are hitting 512 arp entries.
6.0: Known Bugs & RFEs
6.1: RFEs
***Under construction***
6.2: Bugs
***Under construction**7.0: References
====================================
7.1 Glossary:
AUI "Attachment Unit Interface" - the Ethernet/IEEE 802.3 term for
the interface between a MAU and a station. A special kind of
cable known as an "AUI Cable" can attach a MAU to a station at a
distance (up to 50 meters).
MAU "Media Adaptor Unit" - an IEEE 802.3 or Ethernet device which
attaches a station to the cable. Popularly called a
"transceiver". Can be attached by cable to the station or built
into the station.
OSI Reference Model - A model put forth by the ISO for
communication between computer equipment and networks, which
maps out 7 protocol layers.
Top layer: layer number 7: application layer
layer number 6: presentation layer
layer number 5: session layer
layer number 4: transport layer
layer number 3: network layer
layer number 2: data-link layer (e.g. IEEE 802.x)
Bottom layer: layer number 1: physical layer (wire &
electricity)
Repeater - In the "Ethernet" world, a "relay" that regenerates and
cleans up signals, but does no buffering of data packets.
It can extend an Ethernet by strengthening signals, but timing
limitations on Ethernets still limit their size.
Router - A network "relay" that uses a protocol beyond the
data-link protocol to route traffic between LANs and other
network links.
Switched Ethernet - really the same as Ethernet as far as
standards go: acts like a very fast multiport Ethernet bridge
giving an Ethernet to each station. Presumably based on
10BASE-T for most stations.
TCP/IP "Transmission Control Protocol/Internet Protocol" -
literally, two protocols developed for the Defense Data Network
to allow their ARPANET to attach to other networks relatively
transparently. The name also designates the entire family of
protocols built out of IP and TCP. The Internet is based upon
TCP/IP.
Concentrator - a device which allows a number of stations to
be connected to a LAN. In the case of Ethernet, it is
simply a multi-port repeater. In the case of ring networks
like Token Ring and FDDI, it acts as a switch which keeps
the ring intact even if individual devices are unplugged.
Coaxial Cable - any of a number of kinds of electrical
communications cable designed so one conductor is in the center
and the second conductor forms a ring around it. Depending upon
who you talk to, someone might have a specific kind of coaxial
cable in mind. Some well known kinds are various Cable TV
cables, cables used by IBM 327x terminals and ARCNet, and cables
used by Ethernet & IEEE 802.3.
Ethernet - LAN data-link protocol developed by a consortium
of vendors, later standardized as IEEE 802.3 with a few
modifications. For many applications, users have not adopted
all the IEEE 802.3 differences. Ethernet/802.3 now can be
run on two types of coaxial cable as well as multi-mode
fiber and unshielded twisted-pair. "Raw" rate of data
transmission is 10 megabits/second.
Fast Ethernet Alliance - Group of vendors working on a 100Mbps
version of IEEE 802.3. They intend to submit their proposals
for approval by the IEEE for a new set of 802.3 standards called
100BASE-T.
Hub - a nebulous term, typically applied to a multiport repeater
or concentrator consisting of a chassis with slots to be
populated by cards, allowing it to be configured with various
numbers and combinations of LAN ports. Vendors of networking
equipment often also have other types of devices that can be
inserted in the slots such as terminal servers, bridges,
routers, gateways, etc.
bus topology Consists of one central piece of cable to which network
devices are attached. Its advantages are distributed
intelligence, ease of installation and configuration (Sun
Ethernet). If one node goes down, the other nodes on a
bus network are not disturbed.
CSMA/CD Carrier Sense Multiple-Access Collision Detection
(CSMA/CD) is an access method which is how two or more
Sun workstations can share a single channel bus
transmission medium. Ethernet uses this access method.
carrier sense The carrier sense is a mechanism that listens to the line
prior to attempting transmission.
multiple-access Multiple-access is the means by which any node can access
the line at any time instead of waiting for a designated
time slot or token.
collision detection Collision detection intervenes when a collision between
two or more transmissions occurs. Through the collision
detection, a node is able to sense and recover from the
event by randomly scheduling a retransmission of the same
message.
controller Refers to the Ethernet controller in the workstation.
transceiver cable Connects the controller to the transceiver tap on the
coaxial cable or to a Multiplexor (MUX) box.
transceiver box Provides electrical isolation between the coaxial cable
and the transceiver cable.
multiplexor box Enables the attachment of up to eight transceiver cables
to a single transceiver tap into your coaxial cable.
coaxial cable Forms the backbone of the Ethernet system.
router Joins two local networks.
gateway Enables networks using different protocols to
communicate.
Ethernet
7.2: Important Man Pages
==========================================
add_drv(1M) add a new device driver to the system
admintool(1M) system administration with a graphical user interface
arp(7) Address Resolution Protocol
be(7) BigMAC Fast Ethernet device driver
boot(1M) start the system kernel or a standalone program
bufmod(7) STREAMS Buffer Module
devlinks(1M) adds /dev entries for miscellaneous devices and pseudo-devices
dlpi(7) Data Link Provider Interface
driver.conf(4) driver.conf - driver configuration file
eeprom(1M) EEPROM display and load utility
el(7) 3COM 3C503 Ethernet device driver
elink(7) 3COM 3C507 Ethernet device driver
elx(7) 3COM EtherLink III Ethernet device drive
ethers(4) Ethernet address to hostname database or domain
hosts(4) hosts - host name database
icmp(7) Internet Control Message Protocol
ie(7) Intel 82586 Ethernet device driver
iee(7) Intel EtherExpress 16 Ethernet device driver
ifconfig(1M) ifconfig - configure network interface parameters
in.rarpd(1M) DARPA Reverse Address Resolution Protocol
in.routed(1M) routed - network routing daemon
inet(3N) inet, inet_addr, inet_network - Internet address manipulation
ip(7) Internet Protocol
kernel(1M) UNIX system executable file containing basic operating system services
le(7) le, lebuffer, ledma - Am7990 (LANCE) Ethernet device driver
modinfo(1M) display information about loaded kernel modules
monitor(1M) SPARC system PROM monitor
ndd(1M) get and set driver configuration parameters
netconfig(4) network configuration database
netmasks(4)
netstat(1M),
networks(4),
nsswitch.conf(4),
openprom(7) PROM monitor configuration interface
perfmeter(1) display system performance values in multiple dials or strip charts
prtconf(1M) print system configuration
pseudo(4) configuration files for pseudo device drivers
qe(7) QEC/MACE Ethernet device driver
routing(4) system support for packet network routing
sbus(4) configuration files for SBus device drivers
smc(7) SMC 8003/8013/8216 Ethernet device driver
smce(7) SMC 3032/EISA dual-channel Ethernet device driver
snoop(1M) capture and inspect network packets
sys-unconfig(1M) undo a system's configuration
sysbus(4) configuration files for ISA, EISA, and MCA bus device drivers
sysdef(1M) output system definition
sysidtool(1M) sysidtool, sysidnet, sysidnis, sysidsys, sysidroot -system configuration
tcp(7) Internet Transmission Control Protocol
ticlts(7) loopback transport providers
tr(7) IBM 16/4 Token Ring Network Adapter device driver
traceroute 4.x
udp(7) Internet User Datagram Protocol
7.3: Sun SRDBs
*** Under construction ****
7.4: Sun Educational Services
There aren't any specific courses available.
7.5: Solaris Documentation
*** Under construction ****
7.6: Third Party Documentation
RFC:
NETWORK/DATA LINK LAYER RFCs
RFC-826 An Ethernet address resolution protocol
RFC-894 Standard for the transmission of IP datagrams
over Ethernet networks
RFC-903 A Reverse Address Resolution Protocol
RFC-906 * Bootstrap loading using TFTP
RFC-950 * + Internet standard subnetting procedure
RFC-1027 * + Using ARP to implement transparent subnet gateways
RFC-1042 * Standard for the transmission of IP datagrams over
IEEE 802 networks
RFC-1051 * Standard for the transmission of IP datagrams and ARP
packets over ARCNET networks
RFC-1103 * Proposed standard for the transmission of IP datagrams
over FDDI Networks
RFC-1122 * + Requirements for Internet hosts - communication layer
RFC-1127 * + Perspective on the Host Requirements RFCs
RFC-1134 * Point-to-Point Protocol: A proposal for multi-protocol
transmission of datagrams over Point-to-Point links
RFC-1171 * Point-to-Point Protocol for the transmission of
multi-protocol datagrams over Point-to-Point links
RFC-1172 * Point-to-Point Protocol (PPP) initial configuration
options
stb 910
8.0: Supportability
Under Construction
9.0 Additional Support
Under Construction
SOLUTION SUMMARY:
Revision History
작성일자 : 96.08.12
작성자 : 이민호
수정일자 :
수정자 :