[SOLVED]Purgatory: RH9 -> FC1

[Michael Hipp]

Net Llama! wrote:

> On 12/13/2004 05:16 PM, Michael Hipp wrote:
> 
>> Michael Hipp wrote:
>>
>>> To make a long story long ...
>>>
>>> My local server is bad in need of upgrading and I decided to do it by 
>>> the cli/network method rather than using the disks since I have 
>>> several more to do after this and most are some considerable distance 
>>> from here.
>>>
>>> I tried upgrading RH9 -> FC3 but that appears nigh impossible. So I 
>>> dropped back to a stepwise approach and am at the RH9 - > FC1 step. 
>>> Unfortunately the box I'm upgrading was installed as RH9 with 
>>> "everything".
>>>
>>> I'm using the instructions at:
>>>   http://www.brandonhutchinson.com/Upgrading_Red_Hat_Linux_with_yum.html
>>> which seems pretty straightforward.
>>
>>
>>
>> Well, the upgrade via yum and done entirely logged in via SSH seems to 
>> have worked. The box is now FC1.
>>
>> A recap:
>> 1. Followed the instructions linked above to install yum and 
>> fedora-release.
>> 2. Tried yum upgrade but it failed with dependencies on 
>> redhat-config-securitylevel, cadaver, pine, jdkgcj.
>> 3. I uninstalled those. This forced the uninstall of firstboot also.
>> 4. Then set yum -y upgrade running. The dependency tests, test 
>> transaction, downloading and install took about 6 hours total in about 
>> equal chunks each. It downloaded and installed or upgraded 2505 
>> packages totaling about 1.8GB.
>> 5. I then did find /etc -iname '*rpmsave' to look for any config file 
>> issues to be merged. Found none needing attention - surprise.
>> 6. Did /sbin/chkconfig to see if startup services had been changed. It 
>> needed only minor attention.
>> 7. Did /sbin/poweroff -r now and ran to server room to watch the 
>> console. It seemed to shutdown cleanly and reboot cleanly except for...
>> 8. Noticed that ntpd failed to get a time sync on reboot.
>> 9. Came back in here and logged in via ssh.
>> 10. Checked /var/log/messages about ntpd, said "network unreachable".
>> 11. Investigation found it had no gateway (default route). ***BAD this 
>> means the box was unreachable from the outside - this is a showstopper 
>> for remote upgrades.
>> 12. I added the gateway with 'route add default gw 192.168.0.1'
>> 13. All seems well.
> 
> 
> Did the gateway in /etc/sysconfig/network not get set (properly)?

# cat /etc/sysconfig/networking/devices/ifcfg-eth0
DEVICE=eth0
BROADCAST=192.168.0.255
IPADDR=192.168.0.151
NETMASK=255.255.255.0
NETWORK=192.168.0.0
ONBOOT=yes
TYPE=Ethernet
USERCTL=no
PEERDNS=no
GATEWAY=192.168.0.1

This appears identical to another FC1 box I compared it to. Does it need 
to be listed anywhere else?

Michael