Resque tasks falling off the queue

Resque tasks falling off the queue

I am having a problem where the resque tasks fall off the queue on
deployments. My environment is rails 3.2.11, resque 1.24.1 and redis
2.6.15. The app is deployed on AmazonEC2 instance. I use capistrano for
automated deployments. So here is the issue I am facing in further detail:
Lets say before the deployment I have a few failed tasks in resque queue.
On deploying I loose those entries from the failed queue.
So, every time before deploying I need to ssh into my server go through
any failed or queued tasks make sure they are taken care off and then
deploy. FYI, in my cap script I don't have any hooks called related to
resque. On every deployment I ssh into the current_path and then start
resque-pool daemon through a ssh script.
So, here are my questions:
Is there a way to not affect these queued tasks on every deployment and
just restart the workers?
I have installed redis with apt-get on the Ubuntu server. Is there a
specific setting in /etc/redis/redis.conf that I am missing?
Is there a specific way I need to restart the resque-pool daemon?
Let me know. Thanks in advance.