Red Hat Bugzilla – Bug 23801
Attempted and failed Redhat 7 install on duel processor PowerEdge 4400 w/ RAID 5 (Perc SC)
Last modified: 2007-04-18 12:30:39 EDT
Traceback (innermost last):
File "/usr/bin/anaconda", line 438, in ?
intf.run(todo, test = test)
File "/var/tmp/anaconda-7.0.1//usr/lib/anaconda/text.py", line 1030, in
rc = apply (step(), step)
File "/var/tmp/anaconda-7.0.1//usr/lib/anaconda/text.py", line 507, in
if todo.doInstall ():
File "/usr/lib/anaconda/todo.py", line 1476, in doInstall
File "fstab.py", line 715, in mountFilesystems
raise SystemError, (errno, msg)
SystemError: (6, 'Device not configured')
Local variables in innermost frame:
self: <fstab.NewtFstab instance at 8226138>
msg: Device not configured
Btw, this error came up after an initial error: The installation is proceeding
through the formatting of the partitions I have created. It gets to a certain
partition and then hangs. So I reboot and try to install in text mode, hoping
to shake things up a little. I get to the same point where things hang, and
then I press the magical "Enter" button. The jogs the hang and the formatting
proceeds. After I format some more partitions, I get to the swap partition and
immediately get the error:
"Error creating swap on defvice sda21"
"Error mounting sda19: Device not configured."
If I'm guessing, the first error happened, causing the install to puke. The
program then tried to dump core or write to the tmp/install.log (or whatever
it's called) and then realized that /tmp hadn't been formatted yet, or if it
had, it was all @#$%! up, thus causing the second error message. I believe the
dump listed was from the secondary error and not the originating error.
In total I observed three anomalies:
1. Formatting /q/atheneaum/home/atheneaum6 hangs.
2. Creating a swap file fails.
3. Accessing sda19 fails.
What is the exact hardware specs of the machine you are installing?
Created attachment 7425 [details]
PS - Dell was able to get a copy of Redhat installed. But I didn't like
their "no reasonable partitions" policy. Plus, I wanted to reconfigure the RAID
setup so that I had 1 (of 6) hot swappable backup disks in the RAID array.
After I made this change through the PERC BIOS I attempted to reinstall Redhat.
For the record, I got Linux to install. All I had to do was change all my
18Gb partitions to 38Gb partitions.
I'm guessing there is some implicit limit on the number of partitions that will
format OR that there is some bug concerning the size of the partition name OR a
little bit of both. Most of my partitions were of the form:
/q/atheneaum/home/atheneaum1 (Through 7)
/q/atheneaum/proj/atheneaum1 (Through 6)
All told that's about 20. Is that too many?
Of course I'm not very happy with my new partitioning because now it's a lot
harder to backup based on partitions. Oh well, maybe next time.
Hope this helps.
I think the limit imposed by the linux kernel is 16 partitions per SCSI drive.
Well, that would make a lot of sense since I had 7 home 6 proj and 5 other
partiions. This sums to 18, but it started at 5, so make 23. It dies on '21',
which if you subtrace the 5 is 16. Viola. Normally I look for bugs around
powers of 2 as system limit bugs but the starting at 5 threw me.