Hi,

OS: Solaris 8
Hardware: Sun-Fire-V210

Initially only one of it's 4 ethernet interfaces was configured for
the network connectivity. Recently, all types of communication ends up
freezing (ssh, ftp) and the ping replies are intermittent. The cable
connection and the switch ports are okay.

Windows_machine>ping 10.1.1.2 -t

Pinging 10.1.1.2 with 32 bytes of data:

Reply from 10.1.1.2: bytes=32 time<1ms TTL=255
Request timed out.
<snip>
Request timed out.
Reply from 10.1.1.2: bytes=32 time<1ms TTL=255
Request timed out.

Then I wanted to change the interface settings from bge0 to bge1 after
connecting the cable to corresponding ethernet port. I un-"plumb"ed
bge0 before I come up with bge1.

# cat /etc/inet/hosts
10.1.1.1 test1.test.org test1
10.1.1.2 test2.test.org test2

# cat /etc/hostname.bge1
test2.test.org

# cat /etc/netmasks
10.1.1.0 255.255.0.0

# cat /etc/defaultrouter
10.1.1.90

I also edited the following file to update the host information as
"test2" ...

/etc/nodename
/etc/net/ticlts/hosts
/etc/net/ticots/hosts
/etc/net/ticotsord/hosts
/var/crash

Just to check whether the settings are staying or not, I rebooted the
whole system. The "ifconfig -a" doesn't return any IP address as
configured. We can't communicate with the system unless we use this
command:

# ifconfig bge1 10.1.1.2 netmask 255.255.0.0 broadcast 10.1.255.255
plumb up

What am I missing? I might plan for IP multipathing as it started from
Sol 8.


--
Raqueeb Hassan
Bangladesh

did you do a reconfiguration boot?

reboot -- -r

On Mar 5, 2:43 am, "Raqueeb Hassan" <widean...@gmail.com> wrote:
> Hi,
>
> OS: Solaris 8
> Hardware: Sun-Fire-V210
>
> Initially only one of it's 4 ethernet interfaces was configured for
> the network connectivity. Recently, all types of communication ends up
> freezing (ssh, ftp) and the ping replies are intermittent. The cable
> connection and the switch ports are okay.
>
> Windows_machine>ping 10.1.1.2 -t
>
> Pinging 10.1.1.2 with 32 bytes of data:
>
> Reply from 10.1.1.2: bytes=32 time<1ms TTL=255
> Request timed out.
> <snip>
> Request timed out.
> Reply from 10.1.1.2: bytes=32 time<1ms TTL=255
> Request timed out.
>
> Then I wanted to change the interface settings from bge0 to bge1 after
> connecting the cable to corresponding ethernet port. I un-"plumb"ed
> bge0 before I come up with bge1.
>
> # cat /etc/inet/hosts
> 10.1.1.1 test1.test.org test1
> 10.1.1.2 test2.test.org test2
>
> # cat /etc/hostname.bge1
> test2.test.org
>
> # cat /etc/netmasks
> 10.1.1.0 255.255.0.0
>
> # cat /etc/defaultrouter
> 10.1.1.90
>
> I also edited the following file to update the host information as
> "test2" ...
>
> /etc/nodename
> /etc/net/ticlts/hosts
> /etc/net/ticots/hosts
> /etc/net/ticotsord/hosts
> /var/crash
>
> Just to check whether the settings are staying or not, I rebooted the
> whole system. The "ifconfig -a" doesn't return any IP address as
> configured. We can't communicate with the system unless we use this
> command:
>
> # ifconfig bge1 10.1.1.2 netmask 255.255.0.0 broadcast 10.1.255.255
> plumb up
>
> What am I missing? I might plan for IP multipathing as it started from
> Sol 8.
>
> --
> Raqueeb Hassan
> Bangladesh

"Raqueeb Hassan" <wideangle@gmail.com> writes:
> Initially only one of it's 4 ethernet interfaces was configured for
> the network connectivity. Recently, all types of communication ends up
> freezing (ssh, ftp) and the ping replies are intermittent. The cable
> connection and the switch ports are okay.

Please provide "netstat -nr" and "ifconfig -a" output from the Solaris
system and some basic information about your network. (Is the netmask
really 255.255.0.0? Are all the machines on that network configured
that way?)

> Just to check whether the settings are staying or not, I rebooted the
> whole system. The "ifconfig -a" doesn't return any IP address as
> configured. We can't communicate with the system unless we use this
> command:
>
> # ifconfig bge1 10.1.1.2 netmask 255.255.0.0 broadcast 10.1.255.255
> plumb up

That's not going to work. You're telling the system to configure a
non-existent interface, and then plumb it afterwards. That should be:

# ifconfig bge1 plumb 10.1.1.2 netmask 255.255.0.0 \
broadcast 10.1.255.255 up

Or, equivalently, and more concisely:

# ifconfig bge1 plumb 10.1.1.2/16 broadcast + up

--
James Carlson, Solaris Networking <james.d.carlson@sun.com>
Sun Microsystems / 1 Network Drive 71.232W Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677

On Mar 5, 6:55 pm, James Carlson <james.d.carl...@sun.com> wrote:
> "Raqueeb Hassan" <widean...@gmail.com> writes:
> > Initially only one of it's 4 ethernet interfaces was configured for
> > the network connectivity. Recently, all types of communication ends up
> > freezing (ssh, ftp) and the ping replies are intermittent. The cable
> > connection and the switch ports are okay.
>
> Please provide "netstat -nr" and "ifconfig -a" output from the Solaris
> system and some basic information about your network. (Is the netmask
> really 255.255.0.0? Are all the machines on that network configured
> that way?)
>
> > Just to check whether the settings are staying or not, I rebooted the
> > whole system. The "ifconfig -a" doesn't return any IP address as
> > configured. We can't communicate with the system unless we use this
> > command:
>
> > # ifconfig bge1 10.1.1.2 netmask 255.255.0.0 broadcast 10.1.255.255
> > plumb up
>
> That's not going to work. You're telling the system to configure a
> non-existent interface, and then plumb it afterwards. That should be:
>
> # ifconfig bge1 plumb 10.1.1.2 netmask 255.255.0.0 \
> broadcast 10.1.255.255 up
>
> Or, equivalently, and more concisely:
>
> # ifconfig bge1 plumb 10.1.1.2/16 broadcast + up

Thanks for the replies. After the reboot, the ifconfig -a returns

lo0: flags=1000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4> mtu 8232 index
1
inet 127.0.0.1 netmask ff000000
bge1: flags=1000842<BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 3
inet 10.1.1.2 netmask ff000000 broadcast 10.255.255.255
ether x:x:xx:xx:xx:xx
bge2: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500
index 2
inet 192.168.100.3 netmask ffffff00 broadcast 192.168.100.255
ether x:x:xx:xx:xx:xx

Here I configured annex to be connected to bge2. You are right. The
broadcast seems to be a issue here. How do I change that? What should
be the value of /etc/netmasks for network 10.1.1.0?

# netstat -nr

Routing Table: IPv4
Destination Gateway Flags Ref Use Interface
-------------------- -------------------- ----- ----- ------ ---------
192.168.100.0 192.168.100.3 U 1 0 bge2
127.0.0.1 127.0.0.1 UH 3 6 lo0

Here is a problem. It doesn't say anything about network 10.1.1.0 and
how that should be routed. How do I fix that?


TIA

--
Raqueeb Hassan
Bangladesh

"Raqueeb Hassan" <wideangle@gmail.com> writes:
> > # ifconfig bge1 plumb 10.1.1.2 netmask 255.255.0.0 \
> > broadcast 10.1.255.255 up
[...]
> bge1: flags=1000842<BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 3
> inet 10.1.1.2 netmask ff000000 broadcast 10.255.255.255
> ether x:x:xx:xx:xx:xx

So, now you're using a /8 netmask on that network. What's the *right*
netmask?

Note that this interface isn't marked "up." What exactly do you have
in /etc/hostname.bge1?

It's almost certainly not going to run correctly if the netmask issue
isn't resolved and the interface isn't enabled (marked "up").

> bge2: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500
> index 2
> inet 192.168.100.3 netmask ffffff00 broadcast 192.168.100.255
> ether x:x:xx:xx:xx:xx
>
> Here I configured annex to be connected to bge2. You are right. The
> broadcast seems to be a issue here. How do I change that? What should
> be the value of /etc/netmasks for network 10.1.1.0?

I can't answer that question. This is something that has to come from
*you* or your local network administrator. Netmasks are a property of
the local network configuration. They're not something that can be
deduced from an IP address -- even one that's called a "network."

Plausible masks corresponding to network 10.1.1.0 are /24 through /30,
which correspond to these:

255.255.255.0
255.255.255.128
255.255.255.192
255.255.255.224
255.255.255.240
255.255.255.248
255.255.255.252

> # netstat -nr
>
> Routing Table: IPv4
> Destination Gateway Flags Ref Use Interface
> -------------------- -------------------- ----- ----- ------ ---------
> 192.168.100.0 192.168.100.3 U 1 0 bge2
> 127.0.0.1 127.0.0.1 UH 3 6 lo0
>
> Here is a problem. It doesn't say anything about network 10.1.1.0 and
> how that should be routed. How do I fix that?

It's not there because bge1 has been marked "down" administratively.

Most likely, it's due to a bad entry in /etc/hostname.bge1. Probably,
you have multiple lines in that file rather than a single line.

--
James Carlson, Solaris Networking <james.d.carlson@sun.com>
Sun Microsystems / 1 Network Drive 71.232W Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677

Raqueeb Hassan wrote:
> On Mar 5, 6:55 pm, James Carlson <james.d.carl...@sun.com> wrote:
>
>>"Raqueeb Hassan" <widean...@gmail.com> writes:
>>
>>>Initially only one of it's 4 ethernet interfaces was configured for
>>>the network connectivity. Recently, all types of communication ends up
>>>freezing (ssh, ftp) and the ping replies are intermittent. The cable
>>>connection and the switch ports are okay.
>>
>>Please provide "netstat -nr" and "ifconfig -a" output from the Solaris
>>system and some basic information about your network. (Is the netmask
>>really 255.255.0.0? Are all the machines on that network configured
>>that way?)
>>
>>
>>>Just to check whether the settings are staying or not, I rebooted the
>>>whole system. The "ifconfig -a" doesn't return any IP address as
>>>configured. We can't communicate with the system unless we use this
>>>command:
>>
>>># ifconfig bge1 10.1.1.2 netmask 255.255.0.0 broadcast 10.1.255.255
>>>plumb up
>>
>>That's not going to work. You're telling the system to configure a
>>non-existent interface, and then plumb it afterwards. That should be:
>>
>># ifconfig bge1 plumb 10.1.1.2 netmask 255.255.0.0 \
>> broadcast 10.1.255.255 up
>>
>>Or, equivalently, and more concisely:
>>
>># ifconfig bge1 plumb 10.1.1.2/16 broadcast + up
>
>
> Thanks for the replies. After the reboot, the ifconfig -a returns
>
> lo0: flags=1000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4> mtu 8232 index
> 1
> inet 127.0.0.1 netmask ff000000
> bge1: flags=1000842<BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 3
> inet 10.1.1.2 netmask ff000000 broadcast 10.255.255.255
> ether x:x:xx:xx:xx:xx
> bge2: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500
> index 2
> inet 192.168.100.3 netmask ffffff00 broadcast 192.168.100.255
> ether x:x:xx:xx:xx:xx
>
> Here I configured annex to be connected to bge2. You are right. The
> broadcast seems to be a issue here. How do I change that? What should
> be the value of /etc/netmasks for network 10.1.1.0?
>

How many addresses are you using in that address space?

Remember that the netmask is logically "anded" with an IP address to
determine whether the address is "local" or not. There are potentially
2^24-1 addresses in that space. If you are using 10.1.1.0 through
10.1.1.255 then your netmask would be 255.255.255.0. Since that address
space is designated as an RFC-1918 private network you could assign all
2^24-1 addresses and use 255.0.0.0 as a netmask if you wish.

Thanks for the reply. It worked. I had extra blank line in /etc/
hostname.bge1 and /etc/inet/netmasks files.

> > bge1: flags=1000842<BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 3
> > inet 10.1.1.2 netmask ff000000 broadcast 10.255.255.255
> > ether x:x:xx:xx:xx:xx
>
> So, now you're using a /8 netmask on that network. What's the *right*
> netmask?

I'm employing 254 hosts in this subnet, and that should be a
10.1.1.0/24 network with netmask 255.255.255.0. Thinking that network
as Class A (private), I used the default netmask 255.255.0.0
previously. Now, both works.

> Note that this interface isn't marked "up." What exactly do you have
> in /etc/hostname.bge1?
>
> It's almost certainly not going to run correctly if the netmask issue
> isn't resolved and the interface isn't enabled (marked "up").

# cat /etc/netmasks
10.1.1.0 255.255.0.0

<snip>

> > # netstat -nr
>
> > Routing Table: IPv4
> > Destination Gateway Flags Ref Use Interface
> > -------------------- -------------------- ----- ----- ------ ---------
> > 192.168.100.0 192.168.100.3 U 1 0 bge2
> > 127.0.0.1 127.0.0.1 UH 3 6 lo0
>
> > Here is a problem. It doesn't say anything about network 10.1.1.0 and
> > how that should be routed. How do I fix that?
>
> It's not there because bge1 has been marked "down" administratively.

Thanks for that. I didn't mark that previously.

> Most likely, it's due to a bad entry in /etc/hostname.bge1. Probably,
> you have multiple lines in that file rather than a single line.

You are right here. Deleting the blank lines made it work.

# ifconfig -a
bge2: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500
index 3
inet 10.1.1.2 netmask ff000000 broadcast 10.255.255.255
ether x:x:xx:xx:xx:xx

> How many addresses are you using in that address space?
>
> Remember that the netmask is logically "anded" with an IP address to
> determine whether the address is "local" or not. There are potentially
> 2^24-1 addresses in that space. If you are using 10.1.1.0 through
> 10.1.1.255 then your netmask would be 255.255.255.0. Since that address
> space is designated as an RFC-1918 private network you could assign all
> 2^24-1 addresses and use 255.0.0.0 as a netmask if you wish.

Thanks for that.


Thanks all of you once again.


--
Raqueeb Hassan
Bangladesh