For redundancy and fault-tolerance, my objective is to run a pair (or more) of VC servers. No problem there. But some of the scheduled jobs should only run once per scheduled time, so I need a way to "sync" them. The trick is that server pair can share nothing (although they can communicate with each other through any supported means). Ultimately, this may be a feature request (i.e. "exclusive global jobs" that can be defined on multiple VC instances, but are guaranteed to run on only one instance per trigger), but short of that, do you have any recommendations on how to accomplish this?

Did anything come of this question / request? My group would also like to have high availability and job scheduling of visualcron. I am interested how others might have worked with this requirement. I know you could go to lengths to run an independent database and have each job perform some checking before running. I would also be interested in simply covering Active - Standby scenarios for VC servers.

Sorry for the late response. Not sure why you haven't received one earlier.

This seems to be a feature request. I am not sure what level of redundancy you request but most users are happy with having two servers but one inactive in some way. The easiest way is to click on "Stop Server". Then nothing is executed. So, let say you get a problem with server A you just click on "Start server" on server B and it will continue with that schedule.

---

If you take it a step further, and by no means is this exact scenario a suggestion, just food for thought....

Lets say you have 2 identical servers. We're not talking about Microsoft clustering, as frankly I dislike it. This would be internally handled at the service level of VC.

Server1 (Primary): All jobs configured on it, enabled/disabled as necessary, etc...
Server2 (Secondary/Stopped): Sitting there constantly importing the configuration of Server1 every few minutes or on a schedule, that is not dependant on tasks/jobs. It would be more of a core feature that runs as part of the server settings in Server2. Somehow in the Server settings on Server2, configure it to replicate all jobs/tasks/credentials/conditions/etc... like taking a backup on Server1 and importing 'all' on Server2, keeping Server2 in synch while still stopped/not running any jobs. That internal synch from Server1 would only be active while Server2 is stopped. As soon as you start Server2, it would stop synching from Server1 and if/when you wanted to bring Server1 back into the mix as Primary again, you'd backup Server2, stop it, and import on Server1... which would bring Server1 back into production while Server2 resumes it's automatic synching.

I dunno if this is the best way at all, but it's something to consider. Obviously if you have underlying files/folders/scripts etc kept locally and not via UNC path you'd have to replicate those between servers all the time too.

Our Disaster plan right now is to back up our server settings every hour with a date/timestamp and keep 30 days worth. Those get backed up nightly by our enterprise backup solution, and in an 'event', we'd restore the server settings and underlying data drive (scripts/etc..) to a server at our D.R. site, and run from there instead. But there are always potentials for better solutions. Automated solutions are even better, but only after vetting them out in many scenarios.

Brian

We have Visualcron installed on about 40 servers, and got the grand idea to install it on two servers and have them clustered for failover. We were wondering if Visualcron had built in clustering, but as I have read this post and one from 5 years ago, it doesn't seem like it natively does.

Is there any chance of it happening anytime soon?

This seems to be a feature request. I am not sure what level of redundancy you request but most users are happy with having two servers but one inactive in some way. The easiest way is to click on "Stop Server". Then nothing is executed. So, let say you get a problem with server A you just click on "Start server" on server B and it will continue with that schedule.

This is what we do today. We're hoping for a more elegant/automatic solution.

Another poster had mentioned Microsoft clustering, which absolutely not what we're looking for. In a sense, we're looking for "VisualCron clustering", where multiple instances of VC running on different servers (maybe at different data centers) essentially act like one. A 1+1 architecture would probably be fine, but I can also see the benefit for a 1+1+1 architecture or even n+1.

If you decide this is something you want to do, I'd be happy to provide some additional details on the use-case.