Arch Linux

Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines

Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.

REPEAT: Do NOT report bugs for outdated packages!
Tasklist

FS#2362 - network start fails | dhcpcd eth0 works fine...

Attached to Project: Arch Linux
Opened by Josh Kampmeier (quiet) - Sunday, 13 March 2005, 04:46 GMT
Task Type Bug Report
Category Bugtracker
Status Closed
Assigned To No-one
Architecture not specified
Severity Low
Priority Normal
Reported Version 0.7 Wombat
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

I have set up my etc/rc.conf appropriately.

network is included in DAEMONS, eth0="dhcp" is listed and eth0 is added to INTERFACES. I don't know if this is indeed a bug, or just something I misconfigured. I have arch running fine on my notebook and another older E-machine. This particular PC has the following specs:

ASUS A7N8X (nforce2 chipset)
AMD Athlon XP 2600+
512MB DDR PC 2700
80 GB Western Digital HDD
200 GB Western Digistal HDD
Pioneer DVD-Rom
Lite-On CDRW

here is a copy of rc.conf:


#
# /etc/rc.conf - Main Configuration for Arch Linux
#

#
# Localization
#
# HARDWARECLOCK: set to "UTC" or "localtime"
# TIMEZONE: timezones are found in /usr/share/zoneinfo
# KEYMAP: keymaps are found in /usr/share/kbd/keymaps
# CONSOLEFONT: fount in /usr/share/kbd/consolefonts (only needed for non-us)
# USECOLOR: use ANSI color sequences in startup messages
#
HARDWARECLOCK="localtime"
TIMEZONE=America/Chicago
KEYMAP=us
CONSOLEFONT=
USECOLOR="yes"

# Scan for LVM volume groups at startup, required if you use LVM
USELVM="no"

#
# Networking
#
HOSTNAME="neofelis"

#
# Module to load at boot-up (in this order)
# (prefix a module with a ! to disable it)
#
MODULES=(!usbserial !ide-scsi)

#
# Interfaces to start at boot-up (in this order)
# Declare each interface then list in INTERFACES
# (prefix an interface in INTERFACES with a ! to disable it)
#
# Note: to use DHCP, set your interface to be "dhcp" (eth0="dhcp")
#
lo="lo 127.0.0.1"
eth0="dhcp"
#eth0="eth0 192.168.0.2 netmask 255.255.255.0 broadcast 192.168.0.255"
INTERFACES=(lo eth0)

#
# Routes to start at boot-up (in this order)
# Declare each route then list in ROUTES
# (prefix a route in ROUTES with a ! to disable it)
#
gateway="default gw 192.168.0.1"
ROUTES=(!gateway)

#
# Daemons to start at boot-up (in this order)
# (prefix a daemon with a ! to disable it)
#
DAEMONS=(syslog-ng hotplug !pcmcia network netfs crond kdm)

# End of file


Thanks for your help!
This task depends upon

Closed by  arjan timmerman (blaasvis)
Wednesday, 02 November 2005, 09:22 GMT
Reason for closing:  Works for me
Comment by Jan de Groot (JGC) - Sunday, 13 March 2005, 10:31 GMT
Remove the gateway from your configuration, that gets set by DHCP. This could trigger a different exit status of dhcpcd, resulting in a nice red "FAIL" when starting up network.
Comment by Łukasz Fibinger (lucke) - Sunday, 13 March 2005, 13:09 GMT
As far as I can see he doesn't have any route enabled (gateway is prefixed with !). What if you start/stop /etc/rc.d/network manually, Josh?
Comment by Josh Kampmeier (quiet) - Sunday, 13 March 2005, 14:36 GMT
Correct, gateway is prefixed with '!' (but I also tried with commenting out the 'gateway' line with no success.)

running /etc/rc.d/network start does indeed fail... It sits at [ BUSY ] for quite some time.. (~20 seconds) and then pops up with the usual [ FAIL ]

Thanks for the ideas though! :)
Comment by Judd Vinet (judd) - Friday, 18 March 2005, 20:36 GMT
So running "/usr/sbin/dhcpcd eth0" works fine?

Try commenting out the arguments in /etc/conf.d/dhcpcd. Perhaps they're messing things up somehow.

Comment by Josh Kampmeier (quiet) - Thursday, 12 May 2005, 18:58 GMT
nope.. that doesn't work either. the only arguments were "-t 30 -h $HOSTNAME"

I am also having this same exact problem on my notebook after a new install. dhcpcd ethX works fine, network start always fails..

I can't believe I'm having this issue on 2 machines, and no one else is?
Comment by Judd Vinet (judd) - Sunday, 15 May 2005, 18:10 GMT
Very strange problem, since that's all that the network script does.

[snip]
/usr/sbin/dhcpcd $DHCPCD_ARGS $if
[/snip]

I suppose that, as an easy workaround, you could just disable eth0 in your rc.conf and have a "dhcpcd eth0" in your /etc/rc.local.

Doesn't really solve our mystery though...
Comment by Josh Kampmeier (quiet) - Monday, 16 May 2005, 00:43 GMT
I think I figured out what the problem is, at least with the notebook. Like I said, firewire is also seen as an eth device, there is an integrated natsemi NIC, and I use wireless cards(switch between 2 depending on location).

if I have them all listed in rc.conf; i.e. -

eth0="dhcp"
eth1="dhcp"
eth2="dhcp"

the network will fail, because regardless of the setup, at least one of the items will fail to get IP.. i verified this by removing everything except eth1="dhcp" and only left lo and eth1 under the devices(natsemi is seen as eth1) and it did not fail to start that time. So basically, if I let it do dhcp for natsemi, it will work providing I'm connected via cable. In the even I'm using a card, i can just dhcpcd eth2. Not a big deal.

Perhaps it's something similar with my desktop, but I don't have firewire on that. I'll keep searching though, it's not too important there either, because once it's connected, I rarely reboot. Only when the wife needs to do some photo editing or something(but I'm working on that!).

Thanks for the help and suggestions. I'll be sure to post back again if I figure out what's causing it on my desktop.
Comment by arjan timmerman (blaasvis) - Wednesday, 27 July 2005, 09:18 GMT
status ?
Comment by Josh Kampmeier (quiet) - Wednesday, 27 July 2005, 12:42 GMT
The workaround I posted is pretty much the best to get it working... the solution i ended up doing, was to just disable network from starting on boot.. then, depending on how I am connecting (cable or wireless) I can just 'dhcpcd eth#' for the appropriate device...

Perhaps it's not as elegant as it could be... but it's nicer than a big red '[FAILED]' in the boot sequence...

The only way I can think to solve this--I'm no programmer so I don't know how easy/difficult this would be--is... to set up the script that...as long as one of the devices recieves an address, then it reports as DONE... Of course, in computers with multiple NICs that are to be used at once, this could report a false positive.. Anyways, no big deal for me to just dhcpcd the device I want.. thanks for the help guys.

Loading...