Anyone encountered this similar error? Scheduled job run twice. Although, it only shows once in the scheduler log, we can tell it run twice because we have our own logging as well, and data are being logged twice. Below are the other details:
1. This does not happen always, but happens intermittently.
2. No other server or instance running.
3. This is a long running job (3 or 4 hours average).
Anyone who has idea and how you have resolved this problem? I can't seem to find any solutions online although there are similar cases reported.
I hope someone from Adobe Coldfusion support can see and give advice on this as well.
My first thought is that it's related to the length of the job. The job really takes three or four hours? How does the client (CF in this case) know that the job is still in progress? I mean, HTTP doesn't really give you a way to handle that kind of messaging by itself. You're responsible for providing a pretty quick response for every HTTP request. Are you doing that in these cases?
Dave Watts, Eidolon LLC
Thanks for your response. Yes, usual is 3 to 4 hours. Unfortunately, this is an existing schedule already that has been used for quite a long time already. So, rewriting the whole code is not an option. I was thinking of maybe there is some updates/patches from CF, or any CF admin setting we may have missed?
As for how we know if job is in progress or has finished, we have added additional logging mechanisms to do that. And have set also a request timeout limit.
I'm guessing probably not. You can open a bug report with Adobe, but having an HTTP request/response that runs that long is just asking for trouble. HTTP isn't supposed to work that way. You might be able to solve this quickly by going to each scheduled task HTTP endpoint and using CFTHREAD to do the actual work, while returning a simple response right away that basically says "work is starting ...".
Dave Watts, Eidolon LLC
We've had occasional issues where long running CFM's are executed twice (CF11). We see it on a CFM we use to do batch email sends. When the email batch is large enough, i.e. takes longer than about 5 minutes, the CFM will start executing again, and users complain of receiving duplicate emails. I can see the duplicate CFM executing in Fusion Reactor. I thought I had traced it down to IIS recycling the app pool while that CFM was running, but haven't been able to reproduce it.
This is not a scheduled task, BTW.
Did you ever solve your duplicate email issue? We've been having the same problem for the past two months.