We are filing this report per request from Glen Foster:
Description of the problem
While installing (in default mode) RedHat 6.2 on Axi UltraSparc system,
the install process hangs while performing 'post install steps'.
Same problem was seen when we tried 'text mode' option.
Date: Mon, 03 Apr 2000 12:12:15 -0400
From: "Glen A. Foster" <email@example.com>
To: Tuan Le <Tuan.Le@eng.sun.com>
CC: Erik Troan <firstname.lastname@example.org>, sharma@pixels.Eng.Sun.COM,
Subject: Re: (Sparc) Red Hat Linux 6.2 (Zoot)
 Someone from Sun should file all bug reports with
"bugzilla", our defect-tracking mechanism; we will
see what can be done in the near-term so that
perhaps a set of defect fixes can be made available
soon, as I would like your customers to be able to
install (and use!) Red Hat Linux 6.2
Govind, thanks for filing this defect report; now it's trackable. Also, we
have been unsuccessful in reproducing the hang. We have had success with the
following installation configurations, all using a single file system ("/"):
 An "Everything" install (6.2) from FTP media
 An "Everything" install (6.2) from CD
 An "Everything" install (6.1 base) upgraded to 6.2 over NFS
Is there some special configuration you choose (GNOME or KDE workstation,
server, custom install, etc.) that shows the problem (vs the ones above)?
Also, in case there's a hardware difference/problem, here are the pertinent
parts of the AXi loaner system we have here:
(a) 300MHz UltraSPARC IIi CPU
(b) 8.7GB SCSI hard drive
(c) SCSI CD-ROM drive
(d) 3D RAGE II video controller with 4MB RAM
(e) HAPPY MEAL ethernet controller
We have tried on Axi clone (Rave System) which has same h/w configuration as
your system mentioned above. The problem is reproducable on it.
Following are the install options:
1. Installing from local CDROM.
2. Selected "Everything" through custom install option.
3. Disk has two partitions, 256MB swap and rest is allocated to root partition.
this behavior is duplicated for the following conditions:
install using cdrom, and at the network config screen enter:
choose static config (no dhcp) ...
enter ip address ...
accept defaults for network, netmask, broadcast ...
explicitly enter a hostname
explicitly enter the ip of a gateway in a different sub-net from the static ip
explicitly enter the primary dns in a different sub-net from the static ip ...
proceed with the install; during the post-install screen an anaconda traceback
Traceback (innermost last):
File "/usr/lib/python1.5/threading.py", line 376, in __bootstrap self.run()
File "/usr/lib/anaconda/iw/progress.py", line 16, in run rc =
File "/usr/lib/anaconda/todo.py", line 1252, in doInstall
File "/usr/lib/anaconda/todo.py", line 1065, in writeConfiguration
self.writeNetworkConfig ip = self.network.lookupHostname()
File "/usr/lib/anaconda/todo.py", line 531, in writeNetworkConfig ip =
File "/usr/lib/anaconda/todo.py", line 144, in lookupHostname self.gateway)
File "/usr/lib/anaconda/isys.py", line 230, in configNetDevice return
SystemError: (51, 'Network is unreachable')
The actual problem for this defect should be formally changed here. The problem
is that the installer does not gracefully handle the case where a gateway and/or
DNS server cannot be reached to verify a hostname entered in the network
configuration screen (such as gateway and DNS on different subnet from the local
I am seeing the same problem using the Redhat6.2 i386 ISO image. No matter
what configuration I try, I still hang at "Performing post install
configuration". I am using an Abit BE6.0 rev.2 board, 128mb ram, 2gb SCSI
drive partioned with 1900M as "slash" and 141M as "swap".
Reading further (sorry for the extra email and comments), Glen hit it right on
the nose. I am behind a firewall with IP 192.168.1.x, and trying to get DNS
services from 24.0.X.X from beyond the firewall.
I'm having the exact same problem. I can boot from a rescue disk, how do I go
about finishing the install cleanly? Is there some post install script I can
run from the CD?