AWAITING CUSTOMER FEEDBACK ON BELOW:

We are internally discussing feature regarding the ability to link two or more Servers and we would like your feedback. Below are just loose ideas and might change during the discussion, here are our basic ideas so far:

Roles of Server
A Server can have one or more of the following rules:
- Master - do not listen to any slave. Pushes out updates. Optional local execution.
- Slave - listens to a master for updates. Mirror one or more objects (could be single, selected or all Jobs and other objects as well)
- Agent - No scheduler but executes Tasks and shows up in lists as active agent for remote execution

Changes are done in Master and pushed out to Slaves depending on the Sync rules. By default, Master and Slave schedule by themselves but optionally Slave can act as Agent. An Agent is a Slave without scheduling - the Master decides, according to its schedule, when and where (which Agent) to execute on.

Sync
Sync can be done on zero or more global objects or Jobs. Optionally, only some Jobs or objects can be synced. Sync means that current Jobs, with same Id, will be updated by master when updated on Master. There will also be an option to push change to Master. Optionally you can prevent changes on Slave.

Distributed execution
A Job can be set to execute on one or more Servers (Agents). In Job or Task level you can select which servers it should execute on. The Server must be connected as Agent in order to show up.

Load balancing
Based on performance counters and other values like number of running Jobs you can select a number of Servers that it should execute on. When scheduled it will run on one of these servers. Optionally, it can failover to another agent if it fails (agent not connected).

For example, load balancing could be based on Least load, random or alternate

Also, within load balancing you will be able to limit, per Server:
- number of max concurrent Jobs running (queue or skip)
- lock/queue on Connections to prevent multiple connections to SFTP server for example

Failover
When load balancing is in place there can be some failover mechanism as well. For example when Slave is disconnected.

---

This would be a really nice feature indeed. In fact, this is what our business needs to use VisualCron in an environment where we require enterprise-level availability.
To do a good implementation, I think this would require multiple masters with a shared database. because, how else, when one master is not available, could be decided on which agents to run the job? To maintain availability of the database, one could resort to for instance MS-SQL AlwaysOn.

Hi,

a master/slave situation, I'm thinking how this could work...
We use lots of tasks depending on local files.. How can this be configured in a master/slave situation?
Are all task depending on something else skipped in this solution?

Regards
Erik

---

Hi,

a master/slave situation, I'm thinking how this could work...
We use lots of tasks depending on local files.. How can this be configured in a master/slave situation?
Are all task depending on something else skipped in this solution?

Regards
Erik

What do you want to achieve? There are several parts of this idea/implementation. One is on a very high level to have a backup server. One is on a lower level where you can say that Job should run on Server B if server A is not available.

---

Hi,

a master/slave situation, I'm thinking how this could work...
We use lots of tasks depending on local files.. How can this be configured in a master/slave situation?
Are all task depending on something else skipped in this solution?

Regards
Erik

I'm guessing you could use the copy-file tasks to copy the needed files to other servers? That way you could run tasks that are dependent on these files on all your servers, regardless of their role.

I would like to see this implemented as well. All of our jobs are based on UNC paths so this design would be perfect for us to add redundancy to our jobs without needing to cluster.

I never saw this Sticky until now. I had written up a fairly lengthy Feature Request about active/passive/failover outside of Microsoft Clustering because of the desire for 100% uptime (or as close to it as possible) for VisualCron. I think the ideas here are great and thought i'd toss a link to my request here regarding my initial thoughts.

http://www.visualcron.co....aspx?g=posts&t=6604 

Brian

THIS. IS. NEEDED.
topic watched.

Any news on this? Beta available yet?

Any news on this? Beta available yet?

We have been focusing on monitoring first. We will update this thread when we have something. Thank you!

---

WE NEED THIS
topic watched

This would work great with the new MSMQ triggers so you could scale out processing jobs across nodes! Watching topic as well.

+1 for our team. VisualCron is quickly becoming a key piece to our automation layer and is currently the only part that is a manual fail-over. We mantain a separate instance of VC on a warm standby systems and do backup/restore of the production VC data to keep it as fresh as possible. It would be great to have some enterprise redundancy available through the product.

Is there an estimated time when this will be available??

Our company is looking for this functionality and will probably start exploring other solutions.

Thanks for the feedback. We are still planning the features and best ways of implementation.

---

Thanks for the feedback. We are still planning the features and best ways of implementation.

Hi Henrik, is there any update on the progress of your planning for this HA?
Thanks
James

Thanks for the feedback. We are still planning the features and best ways of implementation.

Hi Henrik, is there any update on the progress of your planning for this HA?
Thanks
James

Hi James,

we ares still planning this. We would love to have your feedback on how you expect it to work and what is important for you.

---

Thanks for the feedback. We are still planning the features and best ways of implementation.

And when? In 10 years? 😠 😠 😠

Thanks for the feedback. We are still planning the features and best ways of implementation.

And when? In 10 years? 😠 😠 😠

This is one of the most complex features we have still left to fully develop. One of few differences between VisualCron and competitor that charge 10-1000x more than our product costs. We have had some other prioritizations while we still are gathering information from customers how they want this to work.

Are you a customer today? I would like to link your account to a company (currently you use gmail).

---

I would also like to see the feature come to life. I like the idea of the slave agent model and would work good for our geo-graphic data center design. Right now we only run jobs from a one site to keep things centralized but the ability to send those executions out more locally would be idea. I would request that agents for both window and linux would be awesome. Right now we do lots of SSH remote command jobs that are prone to network disconnect issues. An agent that could be on that server for local execution would solve all these problems.

I would love to have these features available!!! I recently purchased two licenses for VisualCron and plan to setup one instance as a backup. I have a scheduled task to backup the settings daily so that we can restore to our second instance quickly should the need arise. We have many scheduled tasks to handle a plethora of automated tasks, some of which are critical to daily operations. Having a hot spare for VisualCron would ensure maximum up-time.

Please implement these ideas. I'll beta-test for you if needed 😁

---

Robert Adams Jr.
Systems Integration Manager
Information Services
ArnotHealth

I would love to have these features available!!! I recently purchased two licenses for VisualCron and plan to setup one instance as a backup. I have a scheduled task to backup the settings daily so that we can restore to our second instance quickly should the need arise. We have many scheduled tasks to handle a plethora of automated tasks, some of which are critical to daily operations. Having a hot spare for VisualCron would ensure maximum up-time.

Please implement these ideas. I'll beta-test for you if needed 😁

---

Robert Adams Jr.
Systems Integration Manager
Information Services
ArnotHealth

Hi Robert,

please drop an email to support@visualcron.com describing how you would like it to work and what features and settings you would like to see (as detailed as possible).

---

We are looking at using VisualCron for a critical application, which means we have to have some form of High Availability configured for it.

Are there any updates on this master/slave/agent functionality? This sounds infinitely better than trying to use MS clustering. Also we're looking for HA across datacenters, so this would be far superior to trying to stretch a cluster. Plus we can have multiple agents in each datacenter...everything about this method sounds better than the existing clustering option.

When is this feature comming?