I'm getting that message on the log file for mongos.
I have 6 app machines that each have a mongos to connect locally, and that mongos knows how to route to my shard.
Lately I have been getting those messages.
Tue Nov 13 17:14:54 [Balancer] shard0000 is unavailable
Tue Nov 13 17:14:54 [Balancer] shard0000 is unavailable
Tue Nov 13 17:14:54 [Balancer] shard0000 is unavailable
Tue Nov 13 17:14:54 [Balancer] shard0000 is unavailable
Tue Nov 13 17:14:54 [Balancer] shard0000 is unavailable
Tue Nov 13 17:14:54 [Balancer] shard0000 is unavailable
Tue Nov 13 17:14:54 [Balancer] distributed lock 'balancer/myhost-celery-1:27017:1352128252:1804289383' unlocked.
Tue Nov 13 17:14:58 [LockPinger] cluster myhost-mongo-1:27018,myhost-mongo-2:27018,myhost-mongo-3:27018 pinged successfully at Tue Nov 13 17:14:58 2012 by distributed lock pinger 'myhost-mongo-1:27018,myhost-mongo-2:27018,myhost-mongo-3:27018/myhost-celery-1:27017:1352128252:1804289383', sleeping for 30000ms
Tue Nov 13 17:15:18 [Balancer] could not acquire lock 'balancer/myhost-celery-1:27017:1352128252:1804289383' (another update won)
Tue Nov 13 17:15:18 [Balancer] distributed lock 'balancer/myhost-celery-1:27017:1352128252:1804289383' was not acquired.
Tue Nov 13 17:15:31 [Balancer] distributed lock 'balancer/myhost-celery-1:27017:1352128252:1804289383' acquired, ts : 50a28033d02c6ef972249dc0