during install if you look at the temp device nodes that install is creating the device modes c0d11 through c0d15 have the same minor number as c0d1 which means they are all pointing to the c0d1 disk. So install will fail when you try to update the partition table of c0d1 because it is open 7 times. This even occurs with your cciss anaconda update disk. Compaq would like to see this fixed in your next anaconda update.
*** Bug 36099 has been marked as a duplicate of this bug. ***
*** Bug 36100 has been marked as a duplicate of this bug. ***
*** Bug 36101 has been marked as a duplicate of this bug. ***
It looks like you filed the same bug four times with slight variations. Did you mean to file this against 7.0 or 7.1? We have not released an updates disk for anaconda for Red Hat Linux 7.1. 7.1 was just released today (4/16/2001).
These bugs happen with 7.0 and 7.1. I realize that you just released 7.1 yesterday, but we wanted to get this to you so that you could include this when/if you do an anaconda update in the future. This bug was seen on 7.0 and 7.1 and we are checking 6.2. We would like to see it fixed for 7.0 and 7.1. This is not a critical bug since it has been there for at least two releases. We have a workaround, but it is not pretty.
Brock, do we have on of these RAID controllers that use the cciss driver in the test lab?
How is this a kernel bug? The kernel doesn't make the devicefiles, anaconda does.
*** Bug 41870 has been marked as a duplicate of this bug. ***
Can you tell me exactly how to reproduce this bug? I did a test install with a ProLiant with a SmartArray 5300 of Red Hat Linux 7.1 The machine had three drives, which I arranged in a RAID 5 configuration. I created 15 partitions on that RAID device, did the install of 7.1, and then upgraded to our latest internal tree, and eveything worked fine. I must not be doing the same thing that you are...
you only created 15 PARTITIONS you need to create 16 LOGICAL VOLUMES on the array ( d0-d15), then they will see the problem So in your setup create the array, then create 16 logical volumes on that array.
Is this bug still present in 7.1? I have not gotten a clear answer.
oh, I see the bug in the code now. Fixing.
When will we see this fix?
I'm applying the applied patch to CVS. It will be included in the next release. Fixes to prior releases will need to be scheduled.
Created attachment 24187 [details] patch to be applied to CVS
I'm going to move this to resolved, but not closed, since I've fixed in in HEAD. Issues on backporting this can continue to be discussed.