* Milan-re0813.nightly/02-RH-full_server-install-japanese-ftp-TUI.i386 * Null-re0815.1/04-RH-full_workstation-install-german-ftp-TUI.i386 So porkchop has been having a hard week and some of my install cases have running at times when porkchop died. In a couple cases, the installer has simply stalled. There were no tracebacks or stray processes, just nothing happening. In one case, the machine stayed that way for 2 days. In both cases, I see something like the following at the end of anaconda.log: * Unable to find comps.xml * Comps not in update dirs, using ftp://porkchop.test.redhat.com/mnt/redhat/test/Null-re0815.1/i386/i386/RedHat/base/comps.xml * Unable to copy comps package Although I'm not sure if that is related.
Created attachment 70950 [details] anaconda.log
Created attachment 70951 [details] syslog
I'll try to look at its current behaviour.
Has this seemed better with GinGin trees?
This has been better, but I'm not running very many installs from porkchop anymore. I have seen sane retry behavior lately on the few installs that experienced this sort of interruption.
Closing.