Jumpstart client unable to get IP for mac while installation
Hi Techies,
When I initiate boot vnet0 -install from the ok prompt of the client i get the following message contiously...
{0} ok boot vnet1 -v install
Sun Fire T200, No Keyboard
Copyright 2007 Sun Microsystems, Inc. All rights reserved.
OpenBoot 4.26.1, 4096 MB memory available, Serial #66563857.
Ethernet address 0:14:4f:f7:af:11, Host ID: 83f7af11.
Boot device: /virtual-devices@100/channel-devices@200/network@0 File and args: -v install
Requesting Internet Address for 0:14:4f:f7:af:11
Requesting Internet Address for 0:14:4f:f7:af:11
Requesting Internet Address for 0:14:4f:f7:af:11
Requesting Internet Address for 0:14:4f:f7:af:11
Requesting Internet Address for 0:14:4f:f7:af:11
Requesting Internet Address for 0:14:4f:f7:af:11
Requesting Internet Address for 0:14:4f:f7:af:11
On the server when i did a snoop for the MAC - i could see the following
# snoop -d e1000g0 | grep 0:14:4f:f7:af:11
Using device /dev/e1000g0 (promiscuous mode)
OLD-BROADCAST -> (broadcast) RARP C Who is 0:14:4f:f7:af:11 ?
OLD-BROADCAST -> (broadcast) RARP C Who is 0:14:4f:f7:af:11 ?
OLD-BROADCAST -> (broadcast) RARP C Who is 0:14:4f:f7:af:11 ?
OLD-BROADCAST -> (broadcast) RARP C Who is 0:14:4f:f7:af:11 ?
OLD-BROADCAST -> (broadcast) RARP C Who is 0:14:4f:f7:af:11 ?
Server Configuration:
You have entered Hostname = Jumpserv175
IP Address = 20.5.68.175
Terminal Type = vt100
Time Server = localhost
Encrypted Password = 2Scyi05RDsMsY
Name Service = NONE
System Locale = en_US
Network Interface = primary
Netmask = 255.255.255.0
Default Router = 20.5.68.1
IPv6 Used = no
Time Zone = GMT
Security Policy = NONE
Is this correct? (Y/N) [Y]
Modifying add_network_client.sh
Modifying S99jumpstart
Modifying std_svr.profile
Done
Configure the Jumpstart Server
Client Configuration
You have entered Hostname = testdom1
IP Address = 20.5.68.18
MAC Address = 0:14:4f:f7:af:11
Hostid = 83f7af11
Disk Setup = VM50
Veritas Root Mirroring = No
Solaris Root Mirroring = No
ASL Setup = HDS95xx HDS9980k
Kernel Parameters = Standard
Domain Name = xyz.com
SOE Version = 33
Monitor = None
Full Duplex = N
Software Installation = N
Local Customization = None
Begin Script = std_svr.begin
Start Script = std_svr.profile
Finish Script = std_svr.finish
Is this correct? (Y/N) [Y]
Setting up Data Files
Validating rules...
Validating profile Profiles/Solaris_10/std_svr.profile...
The custom JumpStart configuration is ok.
/jumpstart/Scripts/Solaris_10/add_network_client.sh
/jumpstart/Scripts/Solaris_10/add_network_client.sh
Client Addition
# more /etc/bootparams
* install_config=Jumpserv175:/flar
# more /etc/ethers
0:14:4f:f7:af:11 testdom1
Hope these information Suffice... Did I miss any stuff? Can any one help me...
[3160 byte] By [
Vinod_Ka] at [2007-11-27 11:10:16]
# 1
Is that the entire contents of your /etc/ethers ? You would get this error if you have multiple lines in the ethers file that contains the ethernet address, for example:
0:1:2:3:4:5 foo
0:1:2:3:4:5 bar
would make rarpd unable to resolve the address.
Secondly, do you have rarpd running on the jumpstart server? (aka svcs rarp or ps -ef |grep rarp)
.7/M.
# 2
Also check you have the following in OBP on client:
local-mac-address?=false
# 3
You can run in.rarpd in debug mode (with -d).
Also make sure that /etc/ethers is being read. If this is an NIS client, it could have nsswitch.conf set up so that only the NIS map is read, not the local file.
--
Darren
# 4
Hi mAbrante,
I have only one ether entry in /etc/ethers file and i.e
0:14:4f:f7:af:11 testdom1
I have also verified rarp service, it was disabled earlier, but when i enable it - rarp started in maintenance mode and still facing the same error...
# svcs rarp
STATE STIMEFMRI
disabledFeb_09svc:/network/rarp:default
# svcadm enable /network/rarp:default
# svcs rarp
STATE STIMEFMRI
maintenance15:37:38 svc:/network/rarp:default
# svcs -x rarp
svc:/network/rarp:default (Reverse Address Resolution Protocol (RARP) server)
State: maintenance since Mon 19 Feb 2001 03:48:25 PM GMT
Reason: Restarting too quickly.
See: http://sun.com/msg/SMF-8000-L5
See: rarp(7P)
See: in.rarpd(1M)
See: /var/svc/log/network-rarp:default.log
Impact: This service is not running.
*************************************************************************
Hi Stuart_Flisher
This is another good catch Flisher, I have checked the local-mac-address parameter which was 'true' I made it to 'false'.
{0} ok printenv local-mac-address?
local-mac-address? =true
{0} ok setenv local-mac-address? false
local-mac-address? =false
{0} ok printenv local-mac-address?
local-mac-address? =false
We are closer to the fix... that RARP daemon is not running... How to fix?
Vinod
# 5
Disable the service and use Darrens suggestion, aka start it as:
/usr/sbin/in.rarpd -ad
.7/M.
# 6
Hi,
I have tried disabling rarp and enabling it through -ad option as well with svcadm command...
# svcadm disable rarp
# /usr/sbin/in.rarpd -ad
/usr/sbin/in.rarpd: putmsg: Not a stream device
# svcs rarp
STATE STIMEFMRI
disabled17:21:20 svc:/network/rarp:default
# svcadm enable rarp
# svcs rarp
STATE STIMEFMRI
maintenance17:21:51 svc:/network/rarp:default
# /usr/sbin/in.rarpd -ad
/usr/sbin/in.rarpd: putmsg: Not a stream device
# /usr/sbin/in.rarpd -ad vsw0
/usr/sbin/in.rarpd: Usage: /usr/sbin/in.rarpd [ -ad ] device unit
# /usr/sbin/in.rarpd -ad e1000g0
/usr/sbin/in.rarpd: Usage: /usr/sbin/in.rarpd [ -ad ] device unit
An extract of error from "/var/svc/log/network-rarp:default.log"
[ Feb 19 17:21:51 Executing stop method (:kill) ]
[ Feb 19 17:21:51 Executing start method ("/usr/sbin/in.rarpd -a") ]
[ Feb 19 17:21:51 Method "start" exited with status 0 ]
[ Feb 19 17:21:51 Stopping because all processes in service exited. ]
[ Feb 19 17:21:51 Executing stop method (:kill) ]
[ Feb 19 17:21:51 Executing start method ("/usr/sbin/in.rarpd -a") ]
[ Feb 19 17:21:51 Method "start" exited with status 0 ]
[ Feb 19 17:21:51 Stopping because all processes in service exited. ]
[ Feb 19 17:21:51 Executing stop method (:kill) ]
[ Feb 19 17:21:51 Executing start method ("/usr/sbin/in.rarpd -a") ]
[ Feb 19 17:21:51 Method "start" exited with status 0 ]
[ Feb 19 17:21:51 Stopping because all processes in service exited. ]
[ Feb 19 17:21:51 Executing stop method (:kill) ]
[ Feb 19 17:21:51 Restarting too quickly, changing state to maintenance ]
Still the same error... I feel I am missing something...
Vinod
# 7
the -a flag to in.rarpd tells it to bind to all available interfaces, since that failed you could tell it to bind only to e1000g0, you would do that by running:
/usr/sbin/in.rarpd -d e1000g 0
.7/M.
# 8
Hi
# /usr/sbin/in.rarpd -d e1000g 0
/usr/sbin/in.rarpd:[1] device e1000g0 lladdress 0:14:4f:e:bf:ca
/usr/sbin/in.rarpd:[1] device e1000g0 address 0.0.0.0
/usr/sbin/in.rarpd:[1] device e1000g0 subnet mask 0.0.0.0
/usr/sbin/in.rarpd:[3] starting rarp service on device e1000g0 address 0:14:4f:e:bf:ca
# svcs rarp
STATE STIMEFMRI
disabled17:48:01 svc:/network/rarp:default
# ps -ef | grep rarp
root 26336 263180 17:48:12 pts/20:00 grep rarp
root 26314 255740 17:46:09 pts/10:00 /usr/sbin/in.rarpd -d e1000g 0
# svcadm enable rarp
# svcs rarp
STATE STIMEFMRI
maintenance17:48:24 svc:/network/rarp:default
But still rarp daemon is in maintenance mode and jumpstart client is not getting IP for its MAC.
I apologize, I am troubling you a lot and should thank you for providing more and more light towards the solution.
regards,
Vinod
# 9
Well, svcs/svcadm will only affect processes started from the svc.startd process. Since you started in.rarpd manually, it will not be detected by svcs...
According to the output you posted, the e1000g0 interface doesn't seems to have any IP address, is it your primary interface? Perhaps you could post the output of ifconfig -a ?
.7/M.
Message was edited by:
mAbrante
# 10
# ifconfig -a
lo0: flags=2001000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4,VIRTUAL> mtu 8232 index 1
inet 127.0.0.1 netmask ff000000
vsw0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 3
inet 20.5.68.175 netmask ffffff00 broadcast 20.5.68.255
e1000g0: flags=1000842<BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 5
inet 0.0.0.0 netmask 0
ether 0:14:4f:e:bf:ca
For the puropse of LDom creation i have created virtual switch VSW0
Vinod
# 11
why don't you configure interface e1000g0 with a valuable IP address ?
Sakolan.
# 12
Thats a good question - rather
I am creating a guest LDom and this e1000g0 interface is configured as a virtual switch on a primary domain and virtual network port for the guest domains.
I could find in some document that...
vswitch is a layer 2 switch and by default virtual network can't communicate with the external network via the physical interface. As per the LDom install guide, I have unplumbed e1000g0 and plumbed only vsw0. Still the guest unable to get the IP
and i am unable to bind in.rarpd to vsw0...
Vinod
# 13
I can't see the information requested/supplied earlier, but just to cover the simple stuff...
you've got /etc/ethers containing:
<code>
0:14:4f:f7:af:11 testdom1
</code>
What is in /etc/inet/hosts for testdom1 and what does the "hosts" line say in /etc/nsswitch.conf?
# 14
Hi
# grep hosts /etc/nsswitch.conf
hosts:files
# grep testdom1 /etc/hosts
20.5.68.18testdom1
What I could read is the LDom constraints on control and guest domain communication. Need to find a solution for this...
http://blogs.sun.com/bradb/entry/ldoms_for_the_rusty
this above blog has an excerpt of the same problem - but the solution is not clear - if any of you get some clue then it will b of great help for me
"ok boot vnet1 - install
Oh No! It never got it's IP address to start the booting process. What's up with this - I spent alot of time messing around with trying to figure out why rarp wasn't working. Well, if you remember reading the docs the vswitch is a layer 2 switch and by default the vnet can't communicate with the external network via the physical interface. Okay that's cool. I'll just plumb up vsw0 in the control domain. No, it didn't work. The control domain's physical interface (e1000g0) still couldn't see the broadcast from vsw0. Long story short I had to unplumb e1000g0 and plumb up vsw0 per the install guide!!!!"
Vinod
# 15
Hi all,
got the stuffs resolved and thanks for your help...
But its my duty to say how i did it...
Point no 1: LDom has a communication constraint between control and guest domain and vice versa - hence jumpstarting from the control domain to guest domain is not allowed
Point no 2: in.rarpd was not running on the server due to the fact that the rarp couldnot bind to vsw device as there is no device file in /dev directory. hence i have created soft links for vsw and vsw0, thus that problem got solved.
# ln -s /devices/pseudo/clone@0:vsw /dev/vsw
# ln -s /devices/virtual-devices@100/channel-devices@200/virtual-network-switch@0:vsw0 /dev/vsw0
I will proceed with jumpstarting from an external jumpstart server and let you know the updates.,.
Let me assign you all points and bid you off now... Thank you for all your vital information
Vinod
# 16
I accidentally found this thread, it might be useful:
http://forum.java.sun.com/thread.jspa?threadID=5171040
.7/M.
