On one of our publishers there are 4-5 small publications where the data doesn't change too often replicating to a single subscriber. Lately all of them re-try the alloted 10x and write an entry to the application log stating that a deadlock was encountered.
Lets call this publisher A and subscriber A.
Now I realise this does happen every so often since other apps may lock the tables etc.
We have another publisher which replicates all our reference data to all servers. Lets call this publisher B and it replicates to subscriber A too.
When I investigated subscriber A's process's in activity monitor it seems (at random) that replication processes from publisher B are blocking processes from publisher A?
These publications/subscriptions go to the same subscriber database but the actual tables have nothing to do with one another.
Is there a system replication table that is causing a problem or something along those lines?
Publisher C exists in same server location as publisher A and has exact same publications (different content) replication to a different subscriber in same server location as subscriber A.
Publisher B also replicates same data to subscriber C.
None of the above errors are occurring in Pub C to Sub C replication.
Hence I'm confused, what am I missing or overlooking?
View Complete Post