I noticed the following in /var/log/up2date: [Mon Jan 21 19:57:06 2013] up2dateTraceback (most recent call last): File "/usr/sbin/rhn_check", line 332, in __run_action (status, message, data) = CheckCli.__do_call(method, params, kwargs) File "/usr/sbin/rhn_check", line 325, in __do_call retval = method(*params, **kwargs) File "/usr/share/rhn/actions/configfiles.py", line 222, in deploy dep_trans.rollback() File "/usr/share/rhn/config_common/transactions.py", line 262, in rollback os.rmdir(remove_dir)<type 'exceptions.OSError'>: [Errno 39] Directory not empty: '/root'
I'm still not certain why the rollback is necessary or why this even occurs. I think what's happening is that a config deploy fails and it tries to rollback. Because I have files within /root and /home inside the config channels, it tries to delete those directories. /home is empty, so it is removed. /root is not and raises the above exception. Otherwise, why would the kickstart process try to delete /root? I'm still not sure why this is happening. /home is deleted before the kickstart process reaches the %post section. (I checked by putting `ll /home` as the first line in %post.) I suspect it has something to do with the activation key process, as it has some strings attached to it (i.e. join groups, subscribe to config channels, subscribe to child channels). Anyone have ideas on how to troubleshoot? (Spacewalk 1.7 / CentOS 6) Thanks,Giovanni From: [email protected] To: [email protected] Date: Fri, 30 Nov 2012 21:08:54 -0500 Subject: Re: [Spacewalk-list] /home missing after kickstart Thanks, but I'm not certain that is the case. It is happening to all three of my images (CentOS 5, 6, and RHEL 5). It happens when there is combination of using an activation key with subscribing to groups and/or config channels. I need to try a combination of options and see when it happens. Date: Fri, 30 Nov 2012 17:30:55 -0500 From: [email protected] To: [email protected] Subject: Re: [Spacewalk-list] /home missing after kickstart No but I think I can guess what's going on it sounds like the base image anaconda copies to the disk imediatly follow the formatting is corrupt you need to get the boot images off of the cds and reload them On Nov 28, 2012 3:14 PM, "Giovanni Torres" <[email protected]> wrote: I started running into an issue where /home is missing after a kickstart. If I remove everything from the post section, it is still missing. So, that doesn't seem to be the culprit. If I do a blank kickstart, i.e. no activation keys, no groups, no config channels, etc., the /home directory is there. It seems something happens when I try to automatically add an activation key with additional options. The filesystem rpm is installed. I checked the logs in /root after the kickstart and I don't see anything in there that might give an indication why this is happening. All I see is "useradd: cannot create directory /home/<user>" and "mkdir: cannot create directory '/home/dir': No such file or directory. I could add an if statement to the beginning of the post section to create the home directory if it does not already exist, but I feel I shouldn't have to and I'd prefer to find out why this is happening. This happens on both my EL 5 and 6 kickstarts. (I am running Spacewalk 1.7.) Anyone run into this before? Thanks,Giovanni _______________________________________________ Spacewalk-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/spacewalk-list _______________________________________________ Spacewalk-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/spacewalk-list _______________________________________________ Spacewalk-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/spacewalk-list
_______________________________________________ Spacewalk-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/spacewalk-list
