Cron resource breaks on chef-solo version 10.24.0

commented_out.rb

# The resource fails with the commented-out section, but succeeds with the uncommented section:

# Save the nodetool gossipinfo to syslog periodically.  This is to identify                                   
# ghost nodes if we replace one.                                                                              
cron "log_gossip_info" do
  minute "18"
  hour "3,9,15,21"
#   command <<EOF                                                                                             
# nodetool -h0 gossipinfo | awk '{ if($0 ~ /^\//) printf("\n%s", $0) ; else printf("%s", $0) ; } | logger -p \
local1.info -t nodetool_gossipinfo                                                                            
# EOF                                                                                                         
  command <<EOF                                                                                               
nodetool -h0 gossipinfo | logger -p local1.info -t nodetool_gossipinfo                                        
EOF

gistfile1.txt

I'm getting this error:

 * Stopping Tomcat servlet engine tomcat6
   ...done.
 * Starting Tomcat servlet engine tomcat6
   ...done.
[2013-05-13T20:03:07+00:00] INFO: execute[restart_tomcat_if_broken] ran successfully
[2013-05-13T20:03:07+00:00] INFO: Processing execute[restore_cassandra] action run (cassandra::priam_server line 365)
[2013-05-13T20:03:07+00:00] INFO: Processing cron[log_cluster_state] action create (cassandra::priam_server line 376)
[2013-05-13T20:03:08+00:00] INFO: cron[log_cluster_state] added crontab entry
[2013-05-13T20:03:08+00:00] INFO: Processing cron[log_gossip_info] action create (cassandra::priam_server line 386)

================================================================================
Error executing action `create` on resource 'cron[log_gossip_info]'
================================================================================

Chef::Exceptions::Cron
----------------------
Error updating state of log_gossip_info, exit: 1

Resource Declaration:
---------------------
# In /var/tmp/knewton/chef-solo/site-cookbooks/cassandra/recipes/priam_server.rb

386: cron "log_gossip_info" do
387:   minute "18"
388:   hour "3,9,15,21"
389:   command <<EOF
390: nodetool -h0 gossipinfo | awk '{ if($0 ~ /^\//) printf("\n%s", $0) ; else printf("%s", $0) ; } | logger -p local1.info -t nodetool_gossipinfo
391: EOF
392: end
393: 

Compiled Resource:
------------------
# Declared in /var/tmp/knewton/chef-solo/site-cookbooks/cassandra/recipes/priam_server.rb:386:in `from_file'

cron("log_gossip_info") do
  action :create
  retries 0
  retry_delay 2
  minute "18"
  hour "3,9,15,21"
  day "*"
  month "*"
  weekday "*"
  command "nodetool -h0 gossipinfo | awk '{ if($0 ~ /^//) printf(\"\n%s\", $0) ; else printf(\"%s\", $0) ; } | logger -p local1.info -t nodetool_gossipinfo\n"
  user "root"
  cookbook_name :cassandra
  recipe_name "priam_server"
end

[2013-05-13T20:03:08+00:00] ERROR: Running exception handlers
[2013-05-13T20:03:08+00:00] INFO: Exception caught: Log Collection Exception Handler is now running.
[2013-05-13T20:03:08+00:00] INFO: Log tarball_name generated: PN-Cassandra_2013-05-13__20_03
[2013-05-13T20:03:08+00:00] INFO: Collecting the following logs: /var/log/chef/ /var/tmp/knewton/chef-solo/chef-stacktrace.out /var/log/cfn-init.log /var/log/knewton/
[2013-05-13T20:03:08+00:00] INFO: Posting logs to s3://knewton-deploy-failure-logs/PN-Cassandra_2013-05-13__20_03.tar.gz
[2013-05-13T20:03:08+00:00] INFO: Chef run failed, sending CFN signal
[2013-05-13T20:03:08+00:00] INFO: Sending failure signal to CloudFormation for chef[cassandra]@i-402f3023
[2013-05-13T20:03:09+00:00] INFO: CloudFormation signaled successfully with FAILURE.

[2013-05-13T20:03:09+00:00] ERROR: Exception handlers complete
[2013-05-13T20:03:09+00:00] FATAL: Stacktrace dumped to /var/tmp/knewton/chef-solo/chef-stacktrace.out
[2013-05-13T20:03:09+00:00] FATAL: Chef::Exceptions::Cron: cron[log_gossip_info] (cassandra::priam_server line 386) had an error: Chef::Exceptions::Cron: Error updating state of log_gossip_info, exit: 1

And the answer is probably to escape the %. Turns out this wasn't tested/in use yet as I thought.