Hi all,
we have this kind of situation, 1 publicator/distributor and about 40 subscriber. We have about 15 publication and each publication has some articles (from 2 to 15). About half of the publication have filtered articles, the other half haven't filtered articles. We use partitioned publications and pull subscriptions (we run a batch for a portion of the articles every x minutes and another one for the other portion every y minutes). We noticed that if we suspended the synchronization of the non filterd articles the cpu of the server (publicator/distributor) is under control (4-5-10%) and we have no lock, when we include the non filtered articles the cpu is very stressed (about 90-100%) and there are varius lock. Does someone have any ideas to solve this problem? Thanks all!
we have this kind of situation, 1 publicator/distributor and about 40 subscriber. We have about 15 publication and each publication has some articles (from 2 to 15). About half of the publication have filtered articles, the other half haven't filtered articles. We use partitioned publications and pull subscriptions (we run a batch for a portion of the articles every x minutes and another one for the other portion every y minutes). We noticed that if we suspended the synchronization of the non filterd articles the cpu of the server (publicator/distributor) is under control (4-5-10%) and we have no lock, when we include the non filtered articles the cpu is very stressed (about 90-100%) and there are varius lock. Does someone have any ideas to solve this problem? Thanks all!