>Ideally, you should only ever have one process writing to a buffer, and one >process reading (as per the NIST RCS handbook). For multiple processes >reading or writing to a single buffer, it would be better if a >secondary "channel" was spawned for each additional connection - This would >require some small changes to libnml along with the addition of some >arbitration policy.
I did some more checking, and I think only emctask writes to the emcError NML buffer, so it's unlikely that errors get overwritten. There is still the possibility that emctask sends a number of errors, and using no queue would only preserve the last one. I don't think this is very plausible with the current codebase, so Andrew could be fine with this approach. It's pretty hard to test this config change with all possible cases, so I'm a bit reluctant on advising it broadly. Regards, Alex ------------------------------------------------------------------------------ Check out the new SourceForge.net Marketplace. It is the best place to buy or sell services for just about anything Open Source. http://p.sf.net/sfu/Xq1LFB _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
