Hi, when a group in /etc/group contains over 100 members, chgrp stops functioning properly when a _numeric_ group ID is passed as an argument. Note that this is a different situation then the 32 group limit that linux specifies in which a user may not be a part of more than 32 groups. This is when a group contains more than 100 or so users.
Providing a group name as an argument works correctly in all cases. However, when the 100+ member group exists in /etc/group, providing a numeric group ID as an argument sends chgrp into an infinite loop eating up all available resources -- happens to render the 2.4.x OOM killer useless as well. I have recently upgraded from stable to testing, but I've only noticed this bug recently, and I am unsure if it has been there before. * * * RedHat has addressed a similar bug in April (see http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=82619). They have fixed the bug in the RHSA-2003:089-11 advisory (see http://rhn.redhat.com/errata/RHSA-2003-089.html). Is this a known bug in Debian? Please, reply all to this email. It is very important because this is a production machine (for lack of a better name) and provides real services to our users. Having the machine die because a script running from cron hits this bug is not nice on the weekends when no access to the machine can be obtained. Thank you, -Jesse -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]