How to prevent a job starting when another job is running

Introduction

You can use locks to prevent a job starting while another job is running. You can find the information about locks at http://www.sos-berlin.com/doc/en/scheduler.doc/lock.xml

Locks are defined in files in the hot folder (live). Lock file names follow the convention:

the name of the lock followed by .lock.xml.
E.g. lockSample.lock.xml

To use locks, you have to:

declare a lock
assign the lock to the required jobs

Example

A typical lock declaration would be:

    <locks>
        <lock name="lockSample"/>
    </locks>

To assign the lock to your jobs use:

 <?xml version="1.0" encoding="ISO-8859-1"?>
 <job order="yes"
     stop_on_error="no">
    <lock.use lock="lockSample"
              exclusive="yes"/>
    <script language="shell">
        <![CDATA[
 echo "here is the job jobsSample_1"
 echo "I'm not running in parallel with job jobsSample_2"
 ping -n 60 localhost
         ]]>
    </script>
 </job>

 <?xml version="1.0" encoding="ISO-8859-1"?>
 <job order="yes"
     stop_on_error="no">
    <lock.use lock="lockSample" exclusive="yes"/>
    <script language="shell">
        <![CDATA[
 echo "here is the job jobsSample_2"
 echo "I'm not running in parallel with job jobsSample_1"
 ping -n 60 localhost
         ]]>
    </script> 
 </job>

Two job chains using the jobs

 <?xml version="1.0" encoding="ISO-8859-1"?>
 <job_chain>
    <job_chain_node state="100"
                    job="jobsSample_1"
                    next_state="success"
                    error_state="error"/>
    <job_chain_node state="success"/>
    <job_chain_node state="error"/>
 </job_chain>

 <?xml version="1.0" encoding="ISO-8859-1"?>
 <job_chain>
    <job_chain_node state="100"
                    job="jobsSample_2"
                    next_state="success"
                    error_state="error"/>
    <job_chain_node state="success"/>
    <job_chain_node state="error"/>
 </job_chain>

Scope of locks

Where are lock files saved?
- Locks are stored directly in the live folder or any sub-folders.
How do you reference locks in jobs?
- Locks are identified by their path:
  - this is made up of the folder where the lock is stored and of the name of the lock.
- Locks that are located in the same folder as the job can be addressed using the lock's name (omitting the folder).
Example:
- Lock Location
  - Folder live/project_a contains a set of jobs and a lock named my_lock_a
  - Folder live/project_b contains a set of jobs and a lock named my_lock_b
- Lock Usage
  - job_a from the folder named project_a can reference the lock from its folder using for example:
    - <lock.use name="my_lock_a" exclusive="true"/>
  - Job job_b from folder project_b can be configured to use its local lock in a similar manner.
  - Should it be necessary to prevent jobs job_a and job_b from runnig in parallel then they have to use a common lock. This is achieved by referencing a common lock, e.g. job job_b could use a reference to my_lock_a from the folder project_a like this:
    - <lock.use name="/project_a/my_lock_a" exlcusive="yes"/>
  - Caveat: It is possible to use the same lock name in different folders to represent different locks. Jobs using relative lock addressing (i.e. just the locks' name) and omitting the folder name would reference the lock in their local folder. Consider using absolute lock addressing if you want jobs from different folders to make use of the same lock.
  - Hint: Locks that are located directly in the live folder can be addressed with a leading slash as follows:
    - <lock.use name="/my_global_lock" exclusive="yes"/>

Space shortcuts

Page tree

Introduction

Example

Scope of locks