Scheduled tasks stopped working after server reboot

[_whitenoise_]

Bit of a strange one. The server was hard rebooted last night, and since then none of our scheduled tasks will run. They all come back with 'Connection error'. Calling the respective urls from that server using curl works fine, and I've commented out the ip restrictions so I can call them in a browser - again works fine. It's just the scheduled tasks from CF admin that has a problem. Would the scheduler have a problem if it hits a cflocation before the eventual template to execute?

[Guest]

Looks in your server.log in the cf logs.  See if the scheduler service started.  If the lib/neo-?.xml file that holds the tasks got corrupted, they may not be able to load (sorry don't know that off hand).  COuld be the file was being written to at time of the reboot.

Try creating or updating a task, that might fix the file, but if you can see them in CF admin I would imagine the neo-.xml file is fine.

hmmmmmmm, I have seen instance where the scheduler stops working from time to time, and a cf service restart takes care of it.  Maybe try that.

Byron Mann

mannb@hostmysite.com

byronosity@gmail.com

Software Architect

hosting.com | hostmysite.com

http://www.hostmysite.com/?utm_source=bb

[_whitenoise_]

Hi Byron

Thanks for an excellent answer. The service is up and running, and I've tested out creating and running new scheduled tasks that call urls on other servers which all work fine, it just can't do it for it's own. I've been wondering if it's something to do with how CF resolves hosts that's different to how curl / wget on that server does. Do you think that CF has it's own hosts table seperate from the server OS one?

I'll give it a go restarting the CF service. It's a live server which is why I havn't tried this yet, but I reckon it's worth a go. In the mean time I've replicated all the scheduled tasks in cron, so it's not an urgent matter, just an annoying one. Thanks for your help!

[_whitenoise_]

Update: It's started working now. Haven't changed a thing - it's gone and fixed itself overnight. Nothing restarted, nothing modified, it just wasn't working yesterday, and now it does.

I'm drawing up a theory, which may be completely wrong, that CF (or rather, the JVM) caches hosts resolution, and it's now updated it's cache. There's a few sites documenting this:

http://www.talkingtree.com/blog/index.cfm/2006/5/23/Configuring-the-Caching-of-Hostname-Resolution-f...

http://www.dgrigg.com/post.cfm/06/14/2006/Coldfusion-Hostname-Resolution-and-Caching

[Guest]

You're correct java caches the lookup, once the application restarts it clears from memory so no more cache.

Web browsers act the same way,  they do an initial lookup and always use that IP until you restart your browser. I think with Mozilla you can go under the advanced config and set this timeout as well in a similar manner.