Yo, Weâre thinking about using concurrency limits to help solve an issue here on the local cluster. Reading the documentation, it seems a perfect match: This limit is applied globally to all jobs submitted from all schedulers across one HTCondor pool However, later I read that the resources need to be defined in the negotiatorâs configuration file Since we have multiple negotiators here, itâs hard to see how these limits will be in effect globally across the pool. My guess is that this piece of the docs had forgotten that it was assuming Thou Shalt Only Have One Negotiator. I thought it worth checking here : is that correct? That the limits are not pool-wide, they are ânegotiator segment wideâ which in many cases means pool wide, but not always? Or am I missing something? JT |