Spacewalk .4, CentOS 5.2
A cautionary tale o' woe.
I was creating new channels and children for CentOS 5.3, and was trying to save
typing by cloning the child channels. I hadn't realized that cloning copied the
packages in those channels. (There's an enhancement: a clone of a base channel
that copies the child channels.)
So I went to delete the packages... and the system went to its knees,
literally. Apache wasn't serving, and when I got fed up and tried to reboot, it
took over half an hour (watching the clock). Came up - oh, and I noticed that
osad-dispatcher was trying to start before Oracle did) - and when top finally
came up (su'ing to root took literally minutes), it showed java eating 98% of
the CPU. For what may be the second time in a long career, I reniced the java
(+3), and within a minute, everything had returned to normal.
I was getting complaints on the console, so I restarted spacewalk, and
everything's fine. I just deleted the rest of the cloned child channels, and
recreated, and had no problems.
Btw, I'm looking at logs in /var/log, and can't find anything that relates to
what happened. Not tomcat5, not httpd/error_log, messages....
mark
_______________________________________________
Spacewalk-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/spacewalk-list
