On Fri January 11 2008, Jonathan Underwood wrote: > Have a look at the logic used in the fail2ban init file - the start > operation checks for a running fail2ban process. If non is found, it > is safe to delete the lock file. The logic in fail2ban is vulnerable to race conditions, which does not help for pm-utils, where it can easily happen that it is run several times very fast after each other. Btw. the initial problem is to only delete a file at startup. I even found a bug report specifically about this: https://bugzilla.redhat.com/250927 Regards, Till
Description: This is a digitally signed message part.