Failover on the same server?

I read in the documentation that there is a way to configure the master and standby schedulers on the same server. I assumed it is something to do with the installation path so I installed the two instances with different folder names under program files. Unfortunately, that didn't work. It looks like it has something to do with the registry instead. Can you please let me know the steps in order to accomplish what I need? Thanks.

You likely read about running master scheduler and 24x7 Remote Control on the same machine, not a standby scheduler.

As for the master and standby, it is not possible because of the TCP/IP related conflict. You cannot have 2 applications concurrently listening on the same port for network connections. If the reason for running them on the same system is shared file access, you would need to move one 24x7 instance to a local virtual machine. The virtual machine could be configured to run with a different IP address and yet have access to the local files using disk sharing methods.

I thought it wasn't possible looking at the configuration instructions, but the documentation gave me hope (see attached screenshot of documentation).

The reason why I wanted to have this setup is that for some odd reason, our scheduler running as a service becomes unstable after about a week and we would have to restart it. Sometimes we noticed this issue too late and important jobs would not run. I thought having a failover setup would be benefical as a safety net. If you have any insight into this issue, that would be great too. Thanks.[/img]

Apologies, I didn't think it was possible! After reading the help topic text on the screenshot I was able to configure my local instance to start both as a master and as a standby. See my settings on the attached screenshot.

It's good to know it works. I tried setting it up on my side but it's not working quite right. I have the master scheduler installed on "C:\Program Files\24x7 Automation 3" and Standby on "C:\Program Files\24x7 Automation 3_2". Whichever 24x7.exe (master or standby) I used to configure the settings according to your screenshot, both of them will take on the same settings.

In addition, I startup the master scheduler, which opened up fine. Then I attempt to start the standby instance and that's when the error occurs. Please see the screenshots for the 2 messages that I received in sequence from top to bottom. Please let me know what I am doing wrong. Thanks.

The first message is expected. It should appear, but it can be bypassed if you start the standby instance from command line with /NOTIMER option.

As for the second message, I don't get it. How long do you wait after starting the master before trying to start the standby? Does the master complete loading and ready to accept connections from standby? Please post recent entries from your schedule.log

I've waited till the master is completely up and running before I attempted to start the standby. I looked at the schedule.log and noticed an entry about violating the license. I think that's probably the issue. We have 4 standalone licenses and I am guessing the master and standby need different license keys?

The license key is stored in the system registry, they both read it from the same place. There is a built-in license check that prevents networked components to share the same license unless it is a site license. You cannot have master and standby on the same system using different license keys. You would need to have a site license in order to use it with 2 separate installations/instances and allow them to talk to each other.

By the way, why do you want to run master and standby on the same machine? Doesn't it defeat the purpose of a failover in case of machine failure (system shutdown, file corruption, network access problems, etc…)? It is only going to help to deal with scheduler crashes, but there are better ways to deal with them, if such things happen.

Well, since that's the case, then there goes the workaround.

Like I mentioned earlier, the scheduler service would act strange after about a week and the service would need to be restarted. What I mean by strange is I would no longer be able to monitor the jobs using the monitor button from remote control. The screen that appears would be all blank even though I have jobs scheduled to run at certain times throughout the day. Then all jobs would start to error out for no reason at all. After a service restart, then it would be normal again. I noticed that this issue occurs more frequently (sometimes twice a week) when we used global variables so I changed the coding strategy to use script library to return the constants I needed instead. That helped a bit.

Has any other users reported this problem?

Based on what I've heard from you I suspect that the problem is caused by resource leakage happing over time, and I'd like to suggest doing 3 things

1. Make sure all jobs not using global variables are set to run detached. This will ensure that any memory leaks originating in these jobs don't contribute to the memory usage of the scheduler. This should be fairly easy to do and shouldn't affect anything.

2. Use Windows Performance Monitor and set it to monitor memory of the 24x7.exe process with refreshes taking place every 15 to 30 minutes or so. Let it run for a day or 2 and review the chart. This will give you an idea whether the problem is caused leaking memory resources. This also doesn't affect anything.

3. If the Performance Monitor shows such leaks, you would need to update jobs using global variables and make then store global data somewhere else, for example in INI file or registry, not in memory. Appropriate functions for reading writing to files and/or registry are available in 24x7. After that, you can set them to run detached as they would not need to use shared memory.

I believe that steps 1, 2, and 3 described above should correct the issue.