Hi,
I have got a continues transactional replication(push) from SQL 2005 enterprise x64 (sp2) to SQL 2008 enterprise x64 (sp1)
Have built a replication latency monitor, and I have noticed in two situations the replication comes to halt (for the duration of following processes)
When on publisher I run:
1. Defragging indexes (on the tables that are included in replication – bear in mind I am not replicating any indexes)
2. Full database backup (which I use Litespeed 5.0)
I also would like to point out that the distributor host on the publisher instance.
After these processes on publisher replication will catch-up and everything will be in sync.
When I’m running following two statements

I would get IO_COMPLETION wait for sp_replcmds, and my understating is that this command is used by the log reader.
Also defragging indexes will log all the changes in the log file regardless of the recovery method, which put a strain on the replication
What do you think is the best way to take this issue.
Thanks
S