Softpanorama
(slightly skeptical)
Open Source Software Educational Society
May the source be with you, but remember the KISS principle ;-)
Softpanorama Search
|
Softpanorama |
May the source be with you, but remember the KISS principle ;-)
Softpanorama Search
|
Lecture Notes from my FDU class (largely based on Sun Solaris Student Guide materials and O'Reilly books).
Two methods of getting MAC address:
Test question: Which three statements identify information which is displayed by the ifconfig -a command? (Choose three.)
ping command can be used to check and display network connectivity. To use ping the following five conditions must be met:
Changing the hostname: on Solaris the hostname of a system is contained in (at least) six files, which all need to be changed to change the host name of the system ( there can be multiple /etc/hostname.xxn files; one for each networking interface):
Snoop can be used for monitoring network traffic. It switches interface into promiscuous mode. Most important options include:
Note: Press Control-C to stop the snoop utility.
The purpose of this lecture is to give you some understanding of how to set up Solaris network interfaces using the command-line interface. Also included are some basic troubleshooting hints.
The network interfaces that a system uses to communicate with other systems on the network use both hardware and software configuration components. When adding a network interface to a system, you must configure specific files to establish a relationship between the hardware and the software addresses.
Solaris system administrator need to be able:
In order to enable a network interface under Solaris, several steps may be necessary. These include:
An IP address is for each solaris interface is stored in the hostname file, located in the /etc directory. For a system with a single interface (e.g., /dev/bge0), there is one such file. If system has multiple interfaces there are as many files as there are actve interfaces. For example a quad ethernet card (with devices /dev/qfe0, /dev/qfe1, /dev/qfe2, and /dev/qfe3) would have four hostname files containing distinct IP addresses: hostname.qfe0, hostname.qfe1, hostname.qfe2, and hostname.qfe3.
A mulit-homed host allows data to be exchanged only on the local area network (including with the router defined for that network), while a router is responsible for conveying packets between networks. To prevent routing, a multi-homed host must touch the file /etc/notrouter. In addition, the default router for the local network should have its IP address inserted into the file /etc/defaultrouter.
You can create a hosts entry for each interface in the /etc/hosts file or by inserting a record into whatever distributed naming service is mandated by /etc/nsswitch.conf. For example, if the IP address contained in hostname.qfe0, hostname.qfe1, hostname.qfe2, and hostname.qfe3 were to be mapped to the hostnames www1, www2, www3, and www4, the /etc/hosts file would contain the following entries:
# cat /etc/hostswww1 192.64.18.1www2 192.64.18.2www3 192.64.18.3www4 192.64.18.4
Alternatively, if DNS is being used (as shown in Chapter 5), the following entries would need to be made in the appropriate zone file:
www1 IN A 192.64.18.1 ;webserverwww2 IN A 192.64.18.2 ;webserverwww3 IN A 192.64.18.3 ;webserverwww4 IN A 192.64.18.4 ;webserver
Each intarface needs to be plumbed before it can pass and receive
IP traffic.
Once the device is plumbed, its runtime parameters, such as its IP address, can also be configured by using the ifconfig command:
# /usr/sbin/ifconfig hme0 10.64.18.3 broadcast 10.64.18.255 netmask 255.255.255.0
To bring up the interface, the up keyword must be used:
# /usr/sbin/ifconfig hme0 up
All of these individual commands can be combined into the following command, which configures the hardware, sets all parameters, and brings up the interface:
# /usr/sbin/ifconfig hme0 10.64.18.3 broadcast 10.64.18.255 netmask 255.255.255.0 plumb up
In order to determine whether the interfaces are being addressed correctly by other hosts on the local network, use the arp command to display all active connections between the localhost and other hosts:
# /usr/sbin/arp -aNet to Media Table: IPv4Device IP Address Mask Flags Phys Addr------ -------------------- --------------- ----- ---------------hme0 hp 255.255.255.255 00:50:ba:13:08:18hme0 austin 255.255.255.255 SP 00:03:ba:04:a4:e8hme0 224.0.0.0 240.0.0.0 SM 01:00:5e:00:00:00
This displays the ethernet address to IP address mapping for the local host. The flags displayed include:
Once the interface has been enabled, the ifconfig command can be used to view all active interfaces:
# /usr/sbin/ifconfig -alo0: flags=1000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4> mtu 8232 index 1inet 127.0.0.1 netmask ff000000hme0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 2inet 10.64.18.3 netmask ffffff00 broadcast 10.64.18.255lo0: flags=2000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv6> mtu 8252 index 1inet6 ::1/128hme0: flags=2000841<UP,RUNNING,MULTICAST,IPv6> mtu 1500 index 2inet6 fe80::203:baff:fe04:a4e8/10
The media access control (MAC) address is your computer’s unique hardware address on a local area network (LAN). The MAC address is also the Ethernet address on an Ethernet LAN. When you are connected to a LAN, an address resolution table maps your computer’s physical MAC address to an Internet Protocol (IP) address on the LAN. Two ways to display the MAC address or the Ethernet address are:
Note – The MAC address is displayed only if the root user issues the ifconfig command. Only the IP address information is displayed if a non-root user issues the ifconfig command.
You can also retrieve the MAC address from a system that has not yet been booted by performing the banner command at the ok prompt:
ok banner
The ifconfig command displays the current configuration for one of all network (he -a option) interfaces.
When an Ethernet interface is marked as down, it means that it cannot communicate. You can use the ifconfig command to change status of the interface: mark it up or down. For example, to mark the hme0 interface as down use the commands:
# ifconfig hme0 down && ifconfig -a
lo0: flags=1000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4> mtu 8232 index 1
inet 127.0.0.1 netmask ff000000
hme0: flags=1000842<BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 2
inet 192.168.30.41 netmask ffffff00 broadcast 192.168.30.255
ether 8:0:20:93:c9:af
Note – After to take interface down UP flag should no longer be present. When an interface is flagged as UP, it is ready to communicate.
When you mark an interface as up, the UP status appears in the flags field of the ifconfig command output:
# ifconfig hme0 up
# ifconfig -a
lo0: flags=1000849<UP...
The following conditions must be satisfied: for the ping command to succeed in solaris environment:
Snoop utility can capture network packets. This utility is similar to tcpdump and output formats are almost identical. You can use the snoop utility to see what happens when one system tries to communicate with another system. For example the command to view network traffic between two specific systems alpha and beta is:
# snoop alpha, beta
Snoop options include:
Note: Use the -a option to enable audible clicks, which notify you of any network traffic. Although noisy, the clicks are useful when troubleshooting, for example: snoop -a dhcp
Network interfaces in Solaris are controlled by three files:
The /etc/rcS.d/S30network.sh file is one of the startup scripts. The script searches for files called hostname.xxn in the /etc directory, where xx is an interface type and n is the instance of the interface. For every file named /etc/hostname.xxn, the script uses the ifconfig command with the plumb option to make the kernel ready to talk to this type of interface. The script then configures the named interface with an IP address and other required network information.. For example /etc/hostname.hme0 is a file used to configure hme0 interface.
Note: The /etc/rcS.d/S30network.sh file first appeared in the Solaris 8.
The /etc/hostname.xxn file contains an entry for the configuration of a corresponding interface. The xxn component of the file name is replaced by an interface type and a number that differentiates between multiple interfaces of the same type present in the system. For example:
The xx codes for the interface types are product codes. For example, the le code is an abbreviation of the Lance Ethernet, and the qfe code is an abbreviation for Quadfast Ethernet.
The /etc/hostname.hme0 file should contain either the host name or the IP address of the system that contains the hme0 interface. The host name contained in the file must exist in the /etc/hosts file so that it can be resolved to an IP address at system boot time. You can edit the /etc/hostname.hme0 file to contain either the host name or the IP address from the /etc/hosts file, for example:
# cat /etc/hostname.hme0
192.168.1.1
The /etc/inet/hosts file is a local database that associates the IP addresses of hosts with their names. You can use the /etc/inet/hosts file with, or instead of, other hosts databases, including the Domain Name System (DNS), the Network Information Service (NIS) hosts map, and the Network Information Service Plus (NIS+) hosts table. Programs use library interfaces to access information in the /etc/inet/hosts file.
The /etc/inet/hosts file contains at least the loopback and host information. The file has one entry for each IP address of each host. If a host has more than one IP address, this file will have one entry for each address, on separate lines. The format of each line is:
IP-address official-host-name [nicknames] . . .
Items are separated by any number of space or tab characters. The
first item on a line is the host’s IP address. The second entry is the host’s official
name. Subsequent entries on the same line are alternative names for the same machine,
or nicknames. Nicknames are optional. For a host with more than one IP address,
consecutive entries for these
addresses will contain different host names.
# cat /etc/inet/hosts
.
< output truncated>
.
127.0.0.1 localhost
.
< output truncated>
Note – The /etc/inet/hosts file is the official POSIX location of the hosts file. The symbolic link /etc/hosts exists for Berkeley Software Distribution (BSD) compatibility.
You must modify following files to successfully change a system’s hostname:
Editing the /etc/nodename File. Each Solaris OE has a canonical name, which is the official name used when referring to a system. By convention, the system name is the same as the host name associated with the IP address of the primary network interface; for example, hostname.hme0. The following example shows a system’s /etc/nodename file:
# cat /etc/nodename
You can change the canonical name by editing the /etc/nodename file, and rebooting the system. If the machine’s network configuration is managed remotely and delivered by the DHCP or remote procedure calls (RPC) bootparams protocols, the /etc/nodename file is not used. The file is not used because the remote service delivers the canonical name.
Editing the /etc/hostname.xxn File. The /etc/hostname.xxn file contains either the host name or the IP address of the system that contains the named interface.
Editing the /etc/inet/hosts File. Network addresses are written in the conventional decimal-dot notation. Host names are text strings up to 24 characters. Alphabetic characters, numbers, the minus sign, and a period are allowed in the host name. Periods are only allowed when they serve to delimit components of domain style names. Blanks and spaces are not allowed in the host name. No distinction is made between uppercase and lowercase characters. The first character must be an alphabetic character. The last character can't be a minus sign (-) or a dot (.). This file accept regular shell style comments starting with a pound sign (#). All characters after #, up to the end of the line, are not interpreted.
Editing the Three Transport Layer Independent (TLI) Files. The /etc/net directory contains three subdirectories: /etc/net/ticlts, /etc/net/ticots, and /etc/net/ticotsord. Each of these directories contains a hosts file. These files contain configuration information for transport-independent network services. If these files become corrupted, unpredictable results can occur when trying to resolve the system host name when using network services. In addition, when you execute the /usr/sbin/sys-unconfig command, the system deletes all of the hosts files. If the files get corrupted or deleted, you can use any editor to restore them. The format of the file is:
hostname hostname
/usr/sbin/sys-unconfig command can be used to undo a Solaris system configuration. It restores a system’s configuration to an "as-manufactured" state. The system’s configuration includes a host name, NIS domain name, time zone, IP address, IP subnet mask, and root password. The sys-unconfig command does the following:
When the sys-unconfig command is finished, it performs a system shutdown. The sys-unconfig command is a potentially dangerous utility and can only be run by the root user.
When you restart the system, a configuration scripts prompts you to configure the system information. The sys-unconfig command is not available on diskless clients.
Finally, it may be necessary to set some protocol transmission parameters manually to achieve optimal performance. Use the ndd command to set parameters for TCP, UDP, ARP, and IP. In addition, ndd can be used to display the list of all current parameter values relating to a specific protocol. For example, to display the parameters currently associated with TCP, use the following command:
# ndd /dev/tcp \?? (read only)tcp_close_wait_interval (read and write)tcp_conn_req_max_q (read and write)tcp_conn_req_max_q0 (read and write)tcp_conn_req_min (read and write)tcp_conn_grace_period (read and write)tcp_cwnd_max (read and write)tcp_debug (read and write)tcp_smallest_nonpriv_port (read and write)tcp_ip_abort_cinterval (read and write)tcp_ip_abort_linterval (read and write)tcp_ip_abort_interval (read and write)tcp_ip_notify_cinterval (read and write)tcp_ip_notify_interval (read and write)tcp_ip_ttl (read and write)tcp_keepalive_interval (read and write)tcp_maxpsz_multiplier (read and write)tcp_mss_def (read and write)tcp_mss_max (read and write)tcp_mss_min (read and write)tcp_naglim_def (read and write)tcp_rexmit_interval_initial (read and write)tcp_rexmit_interval_max (read and write)tcp_rexmit_interval_min (read and write)tcp_wroff_xtra (read and write)tcp_deferred_ack_interval (read and write)tcp_snd_lowat_fraction (read and write)tcp_sth_rcv_hiwat (read and write)tcp_sth_rcv_lowat (read and write)tcp_dupack_fast_retransmit (read and write)tcp_ignore_path_mtu (read and write)tcp_rcv_push_wait (read and write)tcp_smallest_anon_port (read and write)tcp_largest_anon_port (read and write)tcp_xmit_hiwat (read and write)tcp_xmit_lowat (read and write)tcp_recv_hiwat (read and write)tcp_recv_hiwat_minmss (read and write)tcp_fin_wait_2_flush_interval (read and write)tcp_co_min (read and write)tcp_max_buf (read and write)tcp_zero_win_probesize (read and write)tcp_strong_iss (read and write)tcp_rtt_updates (read and write)tcp_wscale_always (read and write)tcp_tstamp_always (read and write)tcp_tstamp_if_wscale (read and write)tcp_rexmit_interval_extra (read and write)tcp_deferred_acks_max (read and write)tcp_slow_start_after_idle (read and write)tcp_slow_start_initial (read and write)tcp_co_timer_interval (read and write)tcp_extra_priv_ports (read only)tcp_extra_priv_ports_add (write only)tcp_extra_priv_ports_del (write only)tcp_status (read only)tcp_bind_hash (read only)tcp_listen_hash (read only)tcp_conn_hash (read only)tcp_queue_hash (read only)tcp_host_param (read and write)tcp_1948_phrase (write only)
Once all network interfaces are configured as required, use the netstat command, which is responsible for gathering network statistics of various types, to verify their operational status. This data is gathered by using the interfaces on the local host.
netstat is able to gather statistics for the following types of data:
In the following sections, we'll review each of these data gathering operations and discuss how each is used to aid in troubleshooting and pinpointing performance issues.
The per-protocol statistics can be divided into several categories:
RAWIP (raw IP) packets |
TCP packets |
IPv4 packets |
ICMPv4 packets |
IPv6 packets |
ICMPv6 packets |
UDP packets |
IGMP packets |
Each packet type has a specific set of measures associated with it. For example, RAWIP packets have counters that check the number of input (rawipInDatagrams) and output (rawipOutDatagrams) datagrams received since boot. UDP has a corrsponding set of counters that measure the number of input (udpInDatagrams) and output (udpOutDatagrams) datagrams received since boot. In addition to counters of normal events, netstat reports on error events, such as the number of UDP input (udpInErrors) and the number of UDP output (udpOutErrors) errors. These values should be monitored regularly to ensure that the ratio of error to normal conditions does not increase over time. For example, there are 293 tcpActiveOpens shown in the following listing, compared to only one tcpAttemptFails event. If the ratio of tcpAttemptFails to tcpActiveOpens increases over time for TCP traffic, the appropriate TCP parameters may need to be modified by using ndd, or a network error may need to be diagnosed. Here's a representative set of examples for understanding per-protocol errors for IPv6.
$ netstat -sIPv6 ipv6Forwarding = 2 ipv6DefaultHopLimit = 255ipv6InReceives = 0 ipv6InHdrErrors = 0ipv6InTooBigErrors = 0 ipv6InNoRoutes = 0ipv6InAddrErrors = 0 ipv6InUnknownProtos = 0ipv6InTruncatedPkts = 0 ipv6InDiscards = 0ipv6InDelivers = 25 ipv6OutForwDatagrams= 0ipv6OutRequests = 42 ipv6OutDiscards = 2ipv6OutNoRoutes = 0 ipv6OutFragOKs = 0ipv6OutFragFails = 0 ipv6OutFragCreates = 0ipv6ReasmReqds = 0 ipv6ReasmOKs = 0ipv6ReasmFails = 0 ipv6InMcastPkts = 0ipv6OutMcastPkts = 14 ipv6ReasmDuplicates = 0ipv6ReasmPartDups = 0 ipv6ForwProhibits = 0udpInCksumErrs = 0 udpInOverflows = 0rawipInOverflows = 0 ipv6InIPv4 = 0ipv6OutIPv4 = 0 ipv6OutSwitchIPv4 = 0ICMPv6 icmp6InMsgs = 0 icmp6InErrors = 0icmp6InDestUnreachs = 0 icmp6InAdminProhibs = 0icmp6InTimeExcds = 0 icmp6InParmProblems = 0icmp6InPktTooBigs = 0 icmp6InEchos = 0icmp6InEchoReplies = 0 icmp6InRouterSols = 0icmp6InRouterAds = 0 icmp6InNeighborSols = 0icmp6InNeighborAds = 0 icmp6InRedirects = 0icmp6InBadRedirects = 0 icmp6InGroupQueries = 0icmp6InGroupResps = 0 icmp6InGroupReds = 0icmp6InOverflows = 0icmp6OutMsgs = 8 icmp6OutErrors = 0icmp6OutDestUnreachs= 0 icmp6OutAdminProhibs= 0icmp6OutTimeExcds = 0 icmp6OutParmProblems= 0icmp6OutPktTooBigs = 0 icmp6OutEchos = 0icmp6OutEchoReplies = 0 icmp6OutRouterSols = 3icmp6OutRouterAds = 0 icmp6OutNeighborSols= 1icmp6OutNeighborAds = 0 icmp6OutRedirects = 0icmp6OutGroupQueries= 0 icmp6OutGroupResps = 4icmp6OutGroupReds = 0
The per-address statistics can be divided into three categories:
Let's look at sample output from the AF_UNIX sockets:
$ netstat -f unixActive UNIX domain socketsAddress Type Vnode Conn Local Addr Remote Addr30000d03738 stream-ord 30000d1eb78 00000000 /tmp/.X11-unix/X030000d038e0 stream-ord 00000000 0000000030000d03a88 stream-ord 30000ce4a30 00000000 /tmp/jd_sockV630000d03c30 stream-ord 30000a62d78 00000000 /dev/kkcv30000d03dd8 stream-ord 30000a62f50 00000000 /dev/ccv
Here we can see a number of different active sockets using Unix
type addressing, such as the X11 server, which has the address
30000d03738.
The multicast statistics option provides an overview of interfaces
that are currently listening for multicast broadcasts on the
224.0.0.1 (ALL_HOSTS) address.
This is so that packets can be routed appropriately using the router discovery daemon
(in.rdisc), discussed in the next section, "Routing."
In the following example, both the IPv4 and IPv6 multicast groups are displayed:
$ netstat -gGroup Memberships: IPv4Interface Group RefCnt--------- -------------------- ------lo0 224.0.0.1 1hme0 224.0.0.1 1Group Memberships: IPv6If Group RefCnt----- ------------------------ ------lo0 ff02::1:ff00:1 1lo0 ff02::1 1hme0 ff02::202 1hme0 ff02::1:ff04:a4e8 1hme0 ff02::1 2
The kernel maintains a table of routes, constructed by the routing daemon, in.routed. The various routes that have been configured are always viewable by checking the routing statistics:
$ netstat -rRouting Table: IPv4Destination Gateway Flags Ref Use Interface-------------------- -------------------- ----- ----- ------ ---------10.64.18.0 austin U 1 5 hme0224.0.0.0 austin U 1 0 hme0localhost localhost UH 25 215051 lo0
Here, we can see there are two network routes available for packets
on the primary Ethernet interface hme0: the
10.64.18.0 network and the 224.0.0.0
multicast network. In addition, the loopback interface (lo0)
has the local host interface, which is commonly used for troubleshooting and testing.
These routes are all IPv4; however, IPv6 routing details are also displayed:
Routing Table: IPv6Destination/Mask Gateway Flags Ref Use If--------------------------- --------------------------- ----- --- ------ -----fe80::/10 fe80::203:baff:fe04:a4e8 U 1 0 hme0ff00::/8 fe80::203:baff:fe04:a4e8 U 1 0 hme0default fe80::203:baff:fe04:a4e8 U 1 0 hme0localhost localhost UH 5 28 lo0
STREAMS is a System V package that provides access to system calls, standard libraries, and the kernel for the purposes of writing network applications. Any application that uses STREAMS has a specific set of properties about which statistics can be collected, since the I/O operations are distinct from other networking APIs (such as the BSD-style socket API). netstat reports these statistics, including queues, which comprise the read/write operations that characterize a stream:
$ netstat -mstreams allocation:cumulative allocationcurrent maximum total failuresstreams 326 340 7634 0queues 938 962 18662 0mblk 1144 1651 7773 0dblk 1140 1729 2349590 0linkblk 11 169 18 0strevent 9 169 121739 0syncq 25 50 101 0qband 0 0 0 01646 Kbytes allocated for streams data
More details can be obtained by reading the manpage for streamio.
netstat also reports statistics obtained at the IP level. This includes the number of input and output packets counted, the number of input and output errors, and the number of packet collisions. Again, separate entries are shown for IPv4 and IPv6:
$ netstat -iName Mtu Net/Dest Address Ipkts Ierrs Opkts Oerrs Collis Queuelo0 8232 loopback localhost 227695 0 227695 0 0 0hme0 1500 austin austin 2573 0 2130 0 0 0Name Mtu Net/Dest Address Ipkts Ierrs Opkts Oerrs Collislo0 8252 localhost localhost 227705 0 227705 0 0hme0 1500 fe80::203:baff:fe04:a4e8/10 fe80::203:baff:fe04:a4e8 2573 02130 0 0
Most administrators prefer to combine the information that netstat provides into a single report-style format. This can be achieved by using the combined route, socket, and interface statistics, as shown in the output in Example 4-1.
$ netstat -aUDP: IPv4Local Address Remote Address State-------------------- -------------------- -------*.route Idle*.* Unbound*.* Unbound*.sunrpc Idle*.* Unbound*.32771 Idle*.sunrpc Idle*.* Unbound*.32775 Idle*.32779 Idle*.32780 IdleRouting*.* Unbound*.32821 Idle*.32822 Idle*.32823 Idle*.name Idle*.biff Idle*.talk Idle*.time Idle*.echo IdleUDP: IPv6Local Address Remote Address StateIf--------------------------------- --------------------------------- ---------- -----*.* Unbound*.sunrpc Idle*.* Unbound*.32771 Idle*.32779 Idle*.* Unbound*.32821 Idle*.time IdleTCP: IPv4Local Address Remote Address Swind Send-Q Rwind Recv-Q State-------------------- -------------------- ----- ------ ----- ------ -------*.* *.* 0 0 24576 0 IDLE*.sunrpc *.* 0 0 24576 0 LISTEN*.* *.* 0 0 24576 0 IDLE*.sunrpc *.* 0 0 24576 0 LISTEN*.* *.* 0 0 24576 0 IDLE*.32775 *.* 0 0 24576 0 LISTEN*.32776 *.* 0 0 24576 0 LISTEN*.32782 *.* 0 0 24576 0 LISTEN*.32783 *.* 0 0 24576 0 LISTENTCP: IPv6Local Address Remote Address Swind Send-Q Rwind Recv-Q State If*.* *.* 0 24576 0 IDLE*.sunrpc *.* 0 0 24576 0 LISTEN*.* *.* 0 0 24576 0 IDLE*.32775 *.* 0 0 24576 0 LISTENlocalhost.32780 localhost.32775 32768 0 32768 0 CLOSE_WAIT*.32782 *.* 0 0 24576 0 LISTEN*.32791 *.* 0 0 24576 0 LISTEN*.ftp *.* 0 0 24576 0 LISTEN*.telnet *.* 0 0 24576 0 LISTENActive UNIX domain socketsAddress Type Vnode Conn Local Addr Remote Addr30000d03738 stream-ord 30000d1eb78 00000000 /tmp/.X11-unix/X030000d038e0 stream-ord 00000000 0000000030000d03a88 stream-ord 30000ce4a30 00000000 /tmp/jd_sockV630000d03c30 stream-ord 30000a62d78 00000000 /dev/kkcv30000d03dd8 stream-ord 30000a62f50 00000000 /dev/ccv
Some of the TCP messages shown in this output, for both IPv4 and IPv6, may be unfamiliar, so we review each of them individually in Table 4-1.
|
Message |
Description |
|---|---|
|
BOUND |
Socket is bound. |
|
CLOSED |
Socket is closed. |
|
CLOSING |
Socket is closing. |
|
CLOSE_WAIT |
Socket is waiting to close. |
|
ESTABLISHED |
Socket has connected successfully. |
|
FIN_WAIT_1 |
Socket is closing (local). |
|
FIN_WAIT_2 |
Socket is closing (remote). |
|
IDLE |
Socket is idle. |
|
LAST_ACK |
Socket will close after receiving last acknowledgment. |
|
LISTEN |
Socket is active and listening. |
|
SYN_RECEIVED |
Socket is being synchronized. |
|
SYN_SENT |
Socket is creating a connection. |
|
TIME_WAIT |
Socket is waiting to close. |
Imagine that you are a courier, and your run always starts at the local courier depot. You're given a list of addresses, which are associated with a set of packages, and your goal is to deliver them in as little time as possible, subject to the following constraints:
If this seems like a fairly trivial task for a courier, consider how much more difficult the job would be if the following conditions prevailed:
This scenario describes the difficulties faced by the emergence of the Internet and the massive interconnections between hosts and networks. In order for a packet of data to be transferred from host A to host B, a physical path must be identified for the packet to travel.
There is no central lookup service that decides how to route each packet between all possible combinations of two hosts on the Internet (i.e., between the sender and the receiver). This means routes must be generated dynamically. (The only exceptions to this rule are certain situations where a predictable static route may be installed.)
When transferring data around the Internet or between subnets, intermediate hosts must be responsible for transferring packets between networks; these hosts are called routers and are responsible for routing packets between hosts, which can be separated by single subnets or by entire continents. To gain insight into how many routers a packet transfer may involve, let's use the traceroute command to display the "hops" required to connect from a host in Sydney, Australia, to the Sun Microsystems web server:
$ traceroute wwwwseast.usec.sun.com/Tracing route to wwwseast.usec.sun.com [192.9.49.30]over a maximum of 30 hops:1 184 ms 142 ms 138 ms 202.10.4.1312 147 ms 144 ms 138 ms 202.10.4.1293 150 ms 142 ms 144 ms 202.10.1.734 150 ms 144 ms 141 ms atm11-0-0-11.ia4.optus.net.au [202.139.32.17]5 148 ms 143 ms 139 ms 202.139.1.1976 490 ms 489 ms 474 ms hssi9-0-0.sf1.optus.net.au [192.65.89.246]7 526 ms 480 ms 485 ms g-sfd-br-02-f12-0.gn.cwix.net [207.124.109.57]8 494 ms 482 ms 485 ms core7-hssi6-0-0.SanFrancisco.cw.net [204.70.10.9]9 483 ms 489 ms 484 ms corerouter2.SanFrancisco.cw.net [204.70.9.132]10 557 ms 552 ms 561 ms xcore3.Boston.cw.net [204.70.150.81]11 566 ms 572 ms 554 ms sun-micro-system.Boston.cw.net [204.70.179.102]12 577 ms 574 ms 558 ms wwwwseast.usec.sun.com [192.9.49.30]Trace complete.
Here, we can see that some 12 hosts are required to transfer packets between the sender and the receiver. In addition, the observed response times can be quite slow--often more than half a second. It is possible for attempted connections to time out. This can be very useful when trying to identify which intermediate host and/or network is having problems when your remote connection to a host half a world away suddenly dies!
In this section, we'll examine how Solaris solves a number of the classic routing problems.
Static routing typically involves creating a direct physical connection between two hosts, where the implementation of dynamic routing would be wasteful or a security risk. For example, if your local network has three subnets that need to share data, a static route could be created between each router and the other two routers in the network. The number of specific routes required to allow data to flow seamlessly between networks is directly proportional to the square of the number of routers on the network. Every time a change is made to the network topology, these routes will have to be modified manually. If that sounds like too much hard work, consider the situation where it might be desirable: a secure database server that can be accessible only by knowing the route to the host and is not publicly announced. Instead of permitting route discovery, a static route is an appropriate technique here. This could be implemented by creating a point-to-point configuration using ifconfig on a secondary interface, as discussed in the network interface configuration section.
The alternative to static routing is dynamic routing, which involves two daemons: the routing daemon proper (in.routed) and the route discovery daemon (in.rdisc). The in.routed daemon implements the Routing Information Protocol, and is responsible for updating and managing entries in the kernel's routing tables. It uses UDP (port 520) for performing routing operations and operates on all network interfaces that have been plumbed and are identified as up.
If the /etc/notrouter file does not exist, and given that two or more operational interfaces can be found, the host begins to act as a router. Data can then be exchanged between data received on one interface, destined to be transmitted from another interface. For a local area network, the interface that connects to all local hosts is usually known as the internal interface, while the interface that is visible downstream to an ISP or another subnet is known as the external interface. By using packet filtering, it is possible to specify a set of rules governing what type (TCP or UDP) of packets can be transferred between interfaces and on which ports. This is obviously important for protecting local networks, since services that are available to local hosts may not be appropriate for public access.
The route discovery daemon, in.rdisc,
implements the Internet Control Message Protocol (ICMP). In terms of route discovery,
in.rdisc running on host systems listens for multicast
broadcasts on the 224.0.0.1 (ALL_HOSTS)
address. These messages are prioritized, and the default router is selected based
on its proximity to the host. On routers, in.rdisc broadcasts
its availability using multicast on 224.0.0.1, and
listens for requests on 224.0.0.2 (ALL_ROUTERS).
Hosts may request a router directly by broadcasting on 224.0.0.2.
Copyright © 1996-2009 by Dr. Nikolai Bezroukov. www.softpanorama.org was created as a service to the UN Sustainable Development Networking Programme (SDNP) in the author free time. Submit comments This document is an industrial compilation designed and created exclusively for educational use and is placed under the copyright of the Open Content License(OPL). Site uses AdSense so you need to be aware of Google privacy policy. Original materials copyright belong to respective owners. Quotes are made for educational purposes only in compliance with the fair use doctrine.
Disclaimer:
Last modified: August 31, 2009