Michael Olson <[EMAIL PROTECTED]> writes: > One acronym suffices as an explanation of 90% of the delay: AFS.
I don't believe any AFS issues are on the critical path right now, or that any have been during the last week. Please let me know immediately if this is not the case. Also keep in mind that hcoop has just moved from a NIS-style "every machine effectively trusts local root on every other machine" to real distributed security. This is a gigantic architectural change, and it breaks the setuid-corner-cutting most small UNIX sites are used to. This is also why Apache-AFS integration has been the cause of 90% of the AFS-related delay. Apache is up to its ears in code and design choices based on the assumption that "local root" is utterly omnipotent. Getting around that assumption is not easy. - a _______________________________________________ HCoop-Discuss mailing list [email protected] http://hcoop.net/cgi-bin/mailman/listinfo/hcoop-discuss
