Red Hat Bugzilla – Bug 77469
Significant Data-loss type issues on RHN updates
Last modified: 2007-04-18 12:48:12 EDT
From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.1) Gecko/20021003 Description of problem: After initial update through RHN, rpm no longer works properly. It will hang on most install tries. If I install RH8:Custom:Everything, and then *NOT* RHN update, I can function perfectly. If I install RH8:Custom:Everything and run RHN update, I can no longer install NVIDIA drivers, apt-get-rpm, synaptic. (these are the programs I've tried) Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Install RH 8.0 Custom:Everything 2. Run complete RHN update 3. Install NVIDIA, apt-get-rpm, or synaptic (all I tested...It has been part of my standard load for a couple years now) Actual Results: System will run fine if rpm is not used. Any attempt to use rpm from cli : rpm -Uvh, rpm -ivh, etc. will result in the session hanging. You can CTRL-C out of the hang, but ps ax |grep rpm shows the process still alive. You must kill the process at this point. Expected Results: Package should've installed or failed via standard rpm behavior Additional info: I've tried this as a complete install on a P2-333 box with 256MB Ram and 20GB disk twice, and on a Cele 550 with 128M RAM. Reproduces... I have a P3 650 and a P4 1.2 that I will be researching it on today. When it hangs significantly, in some cases it causes a complete hang of the system (no capslock or numlock). On reboot, after recovering journals and cleaning up, KDE3 hangs to reboot in the X load screens.At this point, the system is unusable and requires a reload.
try: rm -rf /var/lib/rpm/__* and trying again. This looks to be a rpm database related locking bug (the __* files are essentially lock files).
Ok... found and fixed problem. Note: If *any* program up2date is going to update is open, or even in memory, the rpm database (instead of not applying the update and giving an error as should be) will simply corrupt itself. a quick edit of inittab to runlevel 3, reboot, rpm --rebuilddb, and then an up2date on a clean boot fixes the problem. Thanks for the help.