Red Hat Bugzilla – Bug 114952
Installer hangs after mkfs
Last modified: 2007-11-30 17:07:00 EST
From Bugzilla Helper:
User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 5.0)
Description of problem:
Anaconda hangs after creating filesystems but before packages are
created. It appears to be trying to read the same installer image
over and over in an endless loop and deadlocks the CPU.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Prepare the VM reader using the kernel & initrd files
3. Enter TCP/IP information as prompted
4. SSH into server as prompted
5. Select either NFS or FTP based install. (Fails on both) Enter
6. Setup partitions
7. Select packages (doesn't seem to matter)
8. SSH a second connection into server to monitor progress (run top)
9. After filesystems are created, but before packages are loaded, the
machine eventually stalls at 100% CPU usage.
Actual Results: Anaconda hangs.
Expected Results: Anaconda installs rpms.
When using NFS server, the NFS server gets saturated with NFS
requests from NFS client that is being installed (running anaconda).
It appears to be requesting the same NFS blocks over and over. If
the NFS server becomes overwhelmed, the NFS client complains about a
compression driver not being able to read blocks.
Platform=s390 (31 bit)
What version are you trying to install? How much memory do you have
for your guests? Can you ssh in again and scp off the /tmp/syslog and
/tmp/anaconda.log and attach them here?
Guest memory was set to 64MB. This was probably the problem. As a
result we set it to 256MB and later dumped the Taroon beta to use the
RHEL3AS Update 1 installer.
All worked well once we set the memory to 256MB.
I suspect that was the root cause.