Cache Digest problems.

I'm having problems with cache digests. Ultimately the digests aren't passing
from peer to peer in my array. Specifics:

3 Squid 2.3.STABLE4 servers compile options enable-cache-digests, enable-poll,
enable-truncate, disable-ident-lookups, enable-heap-replacement, enable-delay-
pools

Squid servers are behind 2 LVS servers for load balancing

squid.conf cache_peer entry

cache_peer titan.inventivecomm.com parent 8080 3130 proxy-only no-query
cache_peer atlas.inventivecomm.com parent 8080 3130 proxy-only no-query
* This is from server 1 pan.inventivecomm.com. Other machines in the array list
the other servers repsectively.

Access.log entry requarding fetch of store_digest.

1 209.155.224.68 CP_MISS/404 370 GET http://atlas.inventivecomm.com:8080/squid-
internal-periodic/store_digest - NONE/- text/plain.

Cachemgr.cgi entry for Peer Selection Algorithms.

peer digest state:
        needed: yes, usable: no, requested: no

        last retry delay: 4800 secs
        last request response time: 1 secs
        last request result: Not Found

peer digest traffic:
        requests sent: 5, volume: 0 KB
        replies recv: 5, volume: 1 KB

peer digest structure:
        no in-memory copy
* Same for all peers

Cachemgr.cgi Store Digest entry

store digest: size: 670 bytes
         entries: count: 907 capacity: 1072 util: 85%
         deletion attempts: 0
         bits: per entry: 5 on: 2630 capacity: 5360 util: 49%
         bit-seq: count: 2656 avg.len: 2.02
         added: 907 rejected: 177 ( 16.33 %) del-ed: 0
         collisions: on add: 1.65 % on rej: 2.26 %

I've read trough the FAQ and the maillist archive and have not found anything
that I see as applicable to this problem. Anyone have a suggestion on what I
might have missed in the configuration to cause this?

Jayme Frye
jfrye@inventivecomm.com

-- 
SysAdmin
Inventive Comm.