Please note that VisualCron support is not actively monitoring this community forum. Please use our contact page for contacting the VisualCron support directly.


Etienne
2012-01-23T16:22:12Z
Hi,

I looked around the forum and did not find anything related to a "Master Server".

When I talk about a Master Server I mean a server that would only be subscribing to other server to monitor their activity.

In our environment where we have tens of servers it impossible to monitor all the job added / removed / unactivated...
Also we can not setup a mail alert for all jobs as we have thousands.

If we could have a server listening to the others and catching alerts, modifications and then send either mail alerts or daily modification summary that would be wonderful.

Does my request is understandable ?

What do you think ?
Sponsor
Forum information
Support
2012-01-24T07:15:10Z
Hi,

we do not have any master server. Currently, a client can connect to many servers and monitor them.

Please explain more about how you would like it to work.
Henrik
Support
http://www.visualcron.com 
Please like  VisualCron on facebook!
Etienne
2012-01-24T13:22:38Z
I'll try to explain my needs as clearly as possible.

We have a quite large environment with thousands of jobs and about 15 person working around the world to manage the jobs and schedules.

One recurrent issue is that someone somewhere add, remove or amend a job and if for some reasons it fails later on or we find it missing it's very very hard to be able to know what was amended or where the job was previously running. So if I could have some kind of report with all the modification applied on all my servers that would make a lot of things easier.

Also when you have a job with multiple tasks and for example one of the task is to run a .bat and the .bat is missing the job will fail with a popup saying the .bat is missing. The issue here is that if no one has a VisualCron running we won't catch the error message. If we had a "Master Server" listening to all the other servers it could for example catch the error popups and automatically send a mail with the job name for example.

Is it a little clearer ?
Support
2012-07-17T18:19:42Z
Hi,

we are currently investigating best way of implementation of a master/slave relationship.

We have some thoughts in mind which we would like to have feedback on. We would also like to have detailed feedback of any other requirements you may have.

Here are our main ideas:

- a VisualCron Server can be configured to be either Master or Slave or Independent
- when a Server is Slave you can specify a Master IP together with a username and password
- the Slave will auto update Jobs, Tasks and other Server objects when the Master is updated
- optionally, the Master can execute Jobs locally
- optionally, the Slaves can be configured to Execute a Job when it is Executed on the Server (manually or by Trigger) - by default Triggers will be located on each Slave machine
- optionally, Settings from Master can be updated on Slaves when updated on Master
- optionally, Jobs can be created on Slaves that is outside of the Master-> sync

We welcome all your feedback regarding this.
Henrik
Support
http://www.visualcron.com 
Please like  VisualCron on facebook!
Etienne
2012-07-18T09:21:52Z
Hi,

Thanks for working on this request.

One of the feature that I have in mind is related to the reporting.

The master should be able to detect all the modifications done on any job so it could generate a report on a pre-defined basis or following a manual request. With this feature we could easily detect a job deletion, a job schedule modification, any trigger loss etc...

Also as all the slaves are connected to the master, the master should be able to detect wrong behavior such as the none execution of job against a planned schedule.

I'll let you know if I think of anything else.
alphadogg
2012-08-01T19:20:37Z
I would like to see this too.

I could use it in a different context in my software shops. They could expand the possibility of using VisualCron in continuous integration. Oftentimes, jobs are parsed out to various servers to distribute building/testing/monitoring loads. There's always a master-salves setup that servers like Jenkins have.

Given the list in the second post by Henrik, I would add that it would be cool if:

  • slaves can have fixed/local jobs
  • slaves can be pushed jobs from master

    • bonus: jobs can be pushed to all slaves, or some slaves, or based on certain conditions, ex: slave is open, random, round-robin, etc.

  • slaves should auto-update based on changes to master, with ability for user to select before push of changes
  • job executions should be able to be farmed out from master to slave

    • bonus: jobs can be pushed to all slaves, or some slaves, or based on certain conditions, ex: slave is available, lowest CPU utilization, random, round-robin, etc.



Hope this helps the design of this feature. Thanks.
Scroll to Top