Network connection diagnostics

Is there a quick or practical way to check a network connection? I
mean, to get IP, gateway address, etc, and to know what is going wrong.
When I just log on, I cannot get to internet, unles I manually issue
ifdown eth0, ifup eth0, service network restart.

Thanks

Marco

---------------------
Freedom is not a permission for chaos.

Marco A. Cruz Quevedo wrote:
> Is there a quick or practical way to check a network connection? I
> mean, to get IP, gateway address, etc, and to know what is going wrong.
> When I just log on, I cannot get to internet, unles I manually issue
> ifdown eth0, ifup eth0, service network restart.
>
> Thanks
>
> Marco

You can activate the hardware from command prompt.
So, there is NOTHING wrong with the network.
I would line up different configuration files and teach _them_ how to
behave.

Have luck

Stanislaw
Slack user from Ulladulla.

Stanislaw Flatto wrote:
> Marco A. Cruz Quevedo wrote:
> > Is there a quick or practical way to check a network connection? I
> > mean, to get IP, gateway address, etc, and to know what is going wrong.
> > When I just log on, I cannot get to internet, unles I manually issue
> > ifdown eth0, ifup eth0, service network restart.
> >
> > Thanks
> >
> > Marco
>
> You can activate the hardware from command prompt.
> So, there is NOTHING wrong with the network.
> I would line up different configuration files and teach _them_ how to
> behave.
>
I have to make all this steps AFTER I was updating some software
packages to my Linux box. The network was ready whe I just logged in,
but now the netowrk does not respond, unless I perform the steps
mentioned above. I know I have to check ALL network related files, so I
can get to know what is wrong, why the network is not ready when I log
into my system.


Regards,

Marco

> Have luck
>
> Stanislaw
> Slack user from Ulladulla.

Followup-To set to comp.os.linux.networking - not applicable to
comp.os.linux.misc - and a stretch for comp.os.linux.setup

On 18 Sep 2006, in the Usenet newsgroup comp.os.linux.networking, in article
<(E-Mail Removed) .com>, Marco A. Cruz Quevedo
wrote:

>Is there a quick or practical way to check a network connection? I
>mean, to get IP, gateway address, etc, and to know what is going wrong.

/sbin/ifconfig -a
/sbin/route -n
cat /etc/resolv.conf

>When I just log on, I cannot get to internet, unles I manually issue
>ifdown eth0, ifup eth0, service network restart.

What distribution - what release? What type of connection to the ISP?

This sounds as if your network configuration scripts are not set correctly,
but there isn't enough information to tell which one[s].

Old guy

On 18 Sep 2006 23:05:50 -0700, "Marco A. Cruz Quevedo"
<(E-Mail Removed)> wrote:

>Is there a quick or practical way to check a network connection? I
>mean, to get IP, gateway address, etc, and to know what is going wrong.
>When I just log on, I cannot get to internet, unles I manually issue
>ifdown eth0, ifup eth0, service network restart.
>
>Thanks
>
>Marco
>
>---------------------
>Freedom is not a permission for chaos.

ifconfig -a shows interfaces (NICs or "adapters"), including their IP
if they have one. route -n shows the routing table, including the
default route if there is one, which default route contains the
gateway address. For example, below, the eth0 NIC's IP is
192.168.3.12. If that line is missing your NIC lacks an IP. And the
gateway address is 192.168.3.2. The default route is the route to
"0.0.0.0" and you see that address in that route's line item in the
table. Maybe your service network restart is triggering an address
request (ie running a dhcp client), but that's not taking place as
part of the boot process. Check /etc/sysconfig/network for something
like ONBOOT=no (or yes). If it's no, it's configured to omit setting
up network stuff as part of booting. Check also for something like
BOOTPROTO=dhcp (I think) to see whether, when the network config is
performed it entails use of dhcp. To experiment with running dhcp
client manually, the name of the client is usually dhclient, otherwise
there's one called pump and one called dhcpcd. Your system might have
one or another of those.

Example:
[root@hostx ~]# ifconfig eth0
eth0 Link encap:Ethernet HWaddr 00:02:B3:41:86:F9
inet addr:192.168.3.12 Bcast:192.168.3.255
Mask:255.255.255.0
inet6 addr: fe80::202:b3ff:fe41:86f9/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:8038 errors:0 dropped:0 overruns:0 frame:0
TX packets:5327 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:10726255 (10.2 MiB) TX bytes:377137 (368.2 KiB)

[root@hostx ~]# route -n
Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use
Iface
192.168.3.0 0.0.0.0 255.255.255.0 U 0 0 0 eth0
0.0.0.0 192.168.3.2 0.0.0.0 UG 0 0 0 eth0
[root@hostx ~]#

Moe Trin wrote:
> Followup-To set to comp.os.linux.networking - not applicable to
> comp.os.linux.misc - and a stretch for comp.os.linux.setup
>
> On 18 Sep 2006, in the Usenet newsgroup comp.os.linux.networking, in article
> <(E-Mail Removed) .com>, Marco A. Cruz Quevedo
> wrote:
>
> >Is there a quick or practical way to check a network connection? I
> >mean, to get IP, gateway address, etc, and to know what is going wrong.
>
> /sbin/ifconfig -a
> /sbin/route -n
> cat /etc/resolv.conf
>
> >When I just log on, I cannot get to internet, unles I manually issue
> >ifdown eth0, ifup eth0, service network restart.
>
> What distribution - what release? What type of connection to the ISP?

Fedora Core 3, connection is DSL
I found that before issuing the commands mentioned above, I can access
internet if I type IP addresses instead of urls. !? Would this be a
clue?

>
> This sounds as if your network configuration scripts are not set correctly,
> but there isn't enough information to tell which one[s].

Perhaps, because this happened after upgrading some RPMS. :-(


>
> Old guy

david wrote:
> On 18 Sep 2006 23:05:50 -0700, "Marco A. Cruz Quevedo"
> <(E-Mail Removed)> wrote:
>
> >Is there a quick or practical way to check a network connection? I
> >mean, to get IP, gateway address, etc, and to know what is going wrong.
> >When I just log on, I cannot get to internet, unles I manually issue
> >ifdown eth0, ifup eth0, service network restart.
> >
> >Thanks
> >
> >Marco
> >
> >---------------------
> >Freedom is not a permission for chaos.
>
> ifconfig -a shows interfaces (NICs or "adapters"), including their IP
> if they have one. route -n shows the routing table, including the
> default route if there is one, which default route contains the
> gateway address. For example, below, the eth0 NIC's IP is
> 192.168.3.12. If that line is missing your NIC lacks an IP. And the
> gateway address is 192.168.3.2. The default route is the route to
> "0.0.0.0" and you see that address in that route's line item in the
> table. Maybe your service network restart is triggering an address
> request (ie running a dhcp client), but that's not taking place as
> part of the boot process. Check /etc/sysconfig/network for something
> like ONBOOT=no (or yes). If it's no, it's configured to omit setting
> up network stuff as part of booting. Check also for something like
> BOOTPROTO=dhcp (I think) to see whether, when the network config is
> performed it entails use of dhcp. To experiment with running dhcp
> client manually, the name of the client is usually dhclient, otherwise
> there's one called pump and one called dhcpcd. Your system might have
> one or another of those.
>
> Example:
> [root@hostx ~]# ifconfig eth0
> eth0 Link encap:Ethernet HWaddr 00:02:B3:41:86:F9
> inet addr:192.168.3.12 Bcast:192.168.3.255
> Mask:255.255.255.0
> inet6 addr: fe80::202:b3ff:fe41:86f9/64 Scope:Link
> UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
> RX packets:8038 errors:0 dropped:0 overruns:0 frame:0
> TX packets:5327 errors:0 dropped:0 overruns:0 carrier:0
> collisions:0 txqueuelen:1000
> RX bytes:10726255 (10.2 MiB) TX bytes:377137 (368.2 KiB)
>
> [root@hostx ~]# route -n
> Kernel IP routing table
> Destination Gateway Genmask Flags Metric Ref Use
> Iface
> 192.168.3.0 0.0.0.0 255.255.255.0 U 0 0 0 eth0
> 0.0.0.0 192.168.3.2 0.0.0.0 UG 0 0 0 eth0
> [root@hostx ~]#

Thanks for the suggestion.

I found that my /etc/resolv.conf is being overwritten with a blank
file! no nameserver addresses in it!

I have not been able to stop this overwriting. It seems that
NetworkManager service is overwiting but this was not happening before.

On Fri, 22 Sep 2006 10:35:38 -0700, Marco A. Cruz Quevedo wrote:

> david wrote:
>> On 18 Sep 2006 23:05:50 -0700, "Marco A. Cruz Quevedo"
>> <(E-Mail Removed)> wrote:
>>
>> >Is there a quick or practical way to check a network connection? I
>> >mean, to get IP, gateway address, etc, and to know what is going wrong.
>> >When I just log on, I cannot get to internet, unles I manually issue
>> >ifdown eth0, ifup eth0, service network restart.
>> >
>> >Thanks
>> >
>> >Marco
>> >
>> >---------------------
>> >Freedom is not a permission for chaos.
>>
>> ifconfig -a shows interfaces (NICs or "adapters"), including their IP
>> if they have one. route -n shows the routing table, including the
>> default route if there is one, which default route contains the
>> gateway address. For example, below, the eth0 NIC's IP is
>> 192.168.3.12. If that line is missing your NIC lacks an IP. And the
>> gateway address is 192.168.3.2. The default route is the route to
>> "0.0.0.0" and you see that address in that route's line item in the
>> table. Maybe your service network restart is triggering an address
>> request (ie running a dhcp client), but that's not taking place as
>> part of the boot process. Check /etc/sysconfig/network for something
>> like ONBOOT=no (or yes). If it's no, it's configured to omit setting
>> up network stuff as part of booting. Check also for something like
>> BOOTPROTO=dhcp (I think) to see whether, when the network config is
>> performed it entails use of dhcp. To experiment with running dhcp
>> client manually, the name of the client is usually dhclient, otherwise
>> there's one called pump and one called dhcpcd. Your system might have
>> one or another of those.
>>
>> Example:
>> [root@hostx ~]# ifconfig eth0
>> eth0 Link encap:Ethernet HWaddr 00:02:B3:41:86:F9
>> inet addr:192.168.3.12 Bcast:192.168.3.255
>> Mask:255.255.255.0
>> inet6 addr: fe80::202:b3ff:fe41:86f9/64 Scope:Link
>> UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
>> RX packets:8038 errors:0 dropped:0 overruns:0 frame:0
>> TX packets:5327 errors:0 dropped:0 overruns:0 carrier:0
>> collisions:0 txqueuelen:1000
>> RX bytes:10726255 (10.2 MiB) TX bytes:377137 (368.2 KiB)
>>
>> [root@hostx ~]# route -n
>> Kernel IP routing table
>> Destination Gateway Genmask Flags Metric Ref Use
>> Iface
>> 192.168.3.0 0.0.0.0 255.255.255.0 U 0 0 0 eth0
>> 0.0.0.0 192.168.3.2 0.0.0.0 UG 0 0 0 eth0
>> [root@hostx ~]#
>
> Thanks for the suggestion.
>
> I found that my /etc/resolv.conf is being overwritten with a blank
> file! no nameserver addresses in it!
>
> I have not been able to stop this overwriting. It seems that
> NetworkManager service is overwiting but this was not happening before.

set the
peerdns=no
in your
/etc/sysconfig/network-scripts/ifcfg-ethX
( or what ever network script your box uses )
and see if it leaves your DNS resolv.conf file
alone.

Most DHCP servers supply accurate info and the
file is overwritten when ever you get a DHCP
update. If it's bad data, you get a bad resolv.conf
file. The peerdns=no should tell your system to stop
updating the file.

jack

--
D.A.M. - Mothers Against Dyslexia

see http://www.jacksnodgrass.com for my contact info.

jack - Grapevine/Richardson

Jack Snodgrass wrote:
> On Fri, 22 Sep 2006 10:35:38 -0700, Marco A. Cruz Quevedo wrote:
>
> > david wrote:
> >> On 18 Sep 2006 23:05:50 -0700, "Marco A. Cruz Quevedo"
> >> <(E-Mail Removed)> wrote:
> >>
> >> >Is there a quick or practical way to check a network connection? I
> >> >mean, to get IP, gateway address, etc, and to know what is going wrong.
> >> >When I just log on, I cannot get to internet, unles I manually issue
> >> >ifdown eth0, ifup eth0, service network restart.
> >> >
> >> >Thanks
> >> >
> >> >Marco
> >> >
> >> >---------------------
> >> >Freedom is not a permission for chaos.
> >>
> >> ifconfig -a shows interfaces (NICs or "adapters"), including their IP
> >> if they have one. route -n shows the routing table, including the
> >> default route if there is one, which default route contains the
> >> gateway address. For example, below, the eth0 NIC's IP is
> >> 192.168.3.12. If that line is missing your NIC lacks an IP. And the
> >> gateway address is 192.168.3.2. The default route is the route to
> >> "0.0.0.0" and you see that address in that route's line item in the
> >> table. Maybe your service network restart is triggering an address
> >> request (ie running a dhcp client), but that's not taking place as
> >> part of the boot process. Check /etc/sysconfig/network for something
> >> like ONBOOT=no (or yes). If it's no, it's configured to omit setting
> >> up network stuff as part of booting. Check also for something like
> >> BOOTPROTO=dhcp (I think) to see whether, when the network config is
> >> performed it entails use of dhcp. To experiment with running dhcp
> >> client manually, the name of the client is usually dhclient, otherwise
> >> there's one called pump and one called dhcpcd. Your system might have
> >> one or another of those.
> >>
> >> Example:
> >> [root@hostx ~]# ifconfig eth0
> >> eth0 Link encap:Ethernet HWaddr 00:02:B3:41:86:F9
> >> inet addr:192.168.3.12 Bcast:192.168.3.255
> >> Mask:255.255.255.0
> >> inet6 addr: fe80::202:b3ff:fe41:86f9/64 Scope:Link
> >> UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
> >> RX packets:8038 errors:0 dropped:0 overruns:0 frame:0
> >> TX packets:5327 errors:0 dropped:0 overruns:0 carrier:0
> >> collisions:0 txqueuelen:1000
> >> RX bytes:10726255 (10.2 MiB) TX bytes:377137 (368.2 KiB)
> >>
> >> [root@hostx ~]# route -n
> >> Kernel IP routing table
> >> Destination Gateway Genmask Flags Metric Ref Use
> >> Iface
> >> 192.168.3.0 0.0.0.0 255.255.255.0 U 0 0 0 eth0
> >> 0.0.0.0 192.168.3.2 0.0.0.0 UG 0 0 0 eth0
> >> [root@hostx ~]#
> >
> > Thanks for the suggestion.
> >
> > I found that my /etc/resolv.conf is being overwritten with a blank
> > file! no nameserver addresses in it!
> >
> > I have not been able to stop this overwriting. It seems that
> > NetworkManager service is overwiting but this was not happening before.
>
> set the
> peerdns=no
> in your
> /etc/sysconfig/network-scripts/ifcfg-ethX
> ( or what ever network script your box uses )
> and see if it leaves your DNS resolv.conf file
> alone.
I have set PEERDNS=no but it still get being overwritten. The only
solution I found is disabling NetworManager service (it seems it is not
being used). I am still looking why this was not happening before.
Anyway, thanks.

Regards,

Marco
>
> Most DHCP servers supply accurate info and the
> file is overwritten when ever you get a DHCP
> update. If it's bad data, you get a bad resolv.conf
> file. The peerdns=no should tell your system to stop
> updating the file.
>
> jack
>
> --
> D.A.M. - Mothers Against Dyslexia
>
> see http://www.jacksnodgrass.com for my contact info.
>
> jack - Grapevine/Richardson

On 22 Sep 2006, in the Usenet newsgroup comp.os.linux.networking, in article
<(E-Mail Removed) .com>, Marco A. Cruz Quevedo
wrote:

>Moe Trin wrote:

>> Marco A. Cruz Quevedo wrote:

>>> When I just log on, I cannot get to internet, unles I manually issue
>>> ifdown eth0, ifup eth0, service network restart.
>>
>> What distribution - what release? What type of connection to the ISP?
>
>Fedora Core 3, connection is DSL
>I found that before issuing the commands mentioned above, I can access
>internet if I type IP addresses instead of urls. !? Would this be a
>clue?

Certainly - as you note on your later post in response to david, this is
a problem with the missing /etc/resolv.conf. The solution proposed by
"Jack Snodgrass" <(E-Mail Removed)> (disabling the
overwriting by the DHCP client) should be one solution, but I'm curious
why the different actions (when booting, the client overwrites with a
blank file, yet when you restart the service, you apparently get the
proper data from the server).

>> This sounds as if your network configuration scripts are not set correctly,
>> but there isn't enough information to tell which one[s].
>
>Perhaps, because this happened after upgrading some RPMS. :-(

That might be the dhcp client package - though I thought they were using
'pump' and the last time I looked at fedoralegacy.org, pump wasn't on the
updates list, and the dhcp*-3.0.1 stuff hasn't been updated since July 2005.

Personally, I'd set the name servers in /etc/resolv.conf, and then set the
DHCP client to leave the file alone. Contrary to the mystique, most ISPs do
not move their name servers around that much, because of the hassle of
getting the word of the new address out to the world. If you've looked at
the DNS-HOWTO, you see that an outsider finds the name server by asking the
root servers, which refer to top-level-domain servers, and they are supposed
to provide the IP addresses of the domain's name servers. You should also be
able to get the same data from a 'whois' query. For example, you seem to be
posting from a Uninet S.A. de C.V address block, and if I ask LACNIC about
this, I am told that the name servers are NSGDL2.UNINET.NET.MX at
200.23.242.204, NSMEX2.UNINET.NET.MX at 200.33.146.204, and
NSMTY2.UNINET.NET.MX at 200.33.148.205, and that information was last changed
in May, when the block 189.128.0.0/10 was allocated by LACNIC. ISPs use DHCP
(Dynamic Host Configuration Protocol) because this makes it simple for the
users - not that the name servers are dancing about.

Old guy