Andrew - Supernews wrote: >>>>>> "W" == W B Hacker <[EMAIL PROTECTED]> writes:
> > No, the problem occurs only if the interface = option is set using > some per-message variable, and isn't the same for every message. > Far more generic an issue than 'interface' alone. That is actually the legacy of the queue architecture and its runners - not originally designed for per-separate-domain or matched-to-interface use. The same issue arises even with multiple, separate configuration files and separate invocations of the (same) binaries. The BFBI fix is to process only one outbound message at a time. A more appropriate fix would require compile-time settings to use separate storage areas - and more. In that case, it is probably less of a maintenance hassle to leave each instance bog-standard and just 'jail' them with fully separate resources and storage. The longer-term fix is to alter the way the queue is structured (per-domain and/or per-interface) and how runners traverse it (restricted to their own tree only). Not a casual item, AFAIK. But I'll be wanting to have a closer at source... Bill -- ## List details at http://www.exim.org/mailman/listinfo/exim-users ## Exim details at http://www.exim.org/ ## Please use the Wiki with this list - http://www.exim.org/eximwiki/
