Alfred von Campe spake the following on 10/2/2006 2:14 PM:

I am in the process of updating a dozen or so workstations from CentOS 4.3 to CentOS 4.4. I set up a local YUM repository, and have been able to successfully update all systems except for one. Both yum and rpm either hang or core dump on this system. I've done a "yum clean all" and an "rpm --rebuilddb" (after deleting the __db files). I was able to successfully update yum itself, but now even a "yum --version" or "rpm -q yum" hang and can only be killed with a "kill -9 <pid>".

Any suggestions on how to proceed? I had no problems on any of the other systems, and they were all configured the same way. I wonder if I may have screwed up when trying to clean up the RPM database (I was following instructions posted a few months ago to this mailing list)...

Alfred

When you updated yum, did you also update sqlite and the other yum dependencies at the same time?

While updating my centOS workstation the other day, I find the update hanging at the line: Updating : udev ##################### [153/398]

I followed the advice of waiting a long time. So I waited at least 2 hours (lunch helped). I know there are always problems interrupting updates.

Finaly, I killed active processes like 10-hal.dev with all kind of parameters (sound, tty, block...) which I could summarize as a huge part of the /dev. There were 2 actives at a time and killing one allowed the next one.

Then, yum update finaly continued updating and finished his job. The reboot went smoothly.

Questions:

1. Am I the only one with this problem?

2. Could I get problems following my killings? I suspect that there were onyl timeouts, but I know nothing about udev.

3. Any idea of what could have been wrong?

4. Are there advices of a better update strategy?

Since centos 4.0, this is my first update problem (4.3 to 4.4).

Anne