README says:
Disque is a synchronously replicated job queue. By default when a new job is added, it is replicated to W nodes before the client gets an acknowledgement about the job being added. W-1 nodes can fail and the message will still be delivered.
What is W by default?
When I add a job without the REPLICATE parameter to a cluster with 3 nodes, job info gives me 'repl 3'. A cluster with 1 node gives me 'repl 1'.
By default REPLICATE=min(cluster-size, 3) or REPLICATE=cluster-size?
This is important because if you don't specify REPLICATE on a cluster with 3 nodes and 1 is down, you won't be able to add jobs.
ADDJOB doc doesn't specify the REPLICATE default value and behavior:
REPLICATE count is the number of nodes the job should be replicated to.
The anchor text in the README is also wrong. Should be addjob instead of 'addjob-queue...'.
Why is the syntax of command and info different?
'REPLICATE 3' in command.
'repl 3' in info. Lowercase and shorter word.
README says:
What is W by default?
When I add a job without the REPLICATE parameter to a cluster with 3 nodes, job info gives me 'repl 3'. A cluster with 1 node gives me 'repl 1'.
By default REPLICATE=min(cluster-size, 3) or REPLICATE=cluster-size?
This is important because if you don't specify REPLICATE on a cluster with 3 nodes and 1 is down, you won't be able to add jobs.
ADDJOB doc doesn't specify the REPLICATE default value and behavior:
The anchor text in the README is also wrong. Should be addjob instead of 'addjob-queue...'.
Why is the syntax of command and info different?
'REPLICATE 3' in command.
'repl 3' in info. Lowercase and shorter word.