CODEWITHSUNDEEP
hadoopyamlansible
dannyman
dannyman

Reputation: 681

In Ansible, how can I access a "number of disks" variable?

I'm looking into tuning MapR Hadoop via Ansible templates.

It is easy enough to tune something to the number of CPU threads found on a system. For example, to set reduce tasks to 1/4 of threads:

<name>mapred.tasktracker.reduce.tasks.maximum</name>
<value>{{ (ansible_processor_vcpus / 4)|int }}</value>

One resource suggests that the number of map / reduce tasks should be scaled to the number of disks on the system. I don't see any comparable variable for that.

There is an array of ansible_devices with sda, sdb, &c. Perhaps I can count that? Perhaps apply a filter so I'm only counting disks which are available to Hadoop?

Upvotes: 0

Views: 518

Answers (1)

Mortimer
Mortimer

Reputation: 2975

maybe length:

<name>mapred.tasktracker.reduce.tasks.maximum</name>
<value>{{ (ansible_devices|length) / 4 }}</value>

Upvotes: 1

Related Questions