I have a local HTTP server running at 127.0.0.1:8888 .. I have 4 tasks set up to run every 10 seconds to hit pages at this server. v5.x worked fine. I upgraded to 6.0.9 last night and it works for 10-30 minutes fine and then HTTP starts timing out (according to VisualCron logs). The only way to resolve the issue is to go in to services.msc and restart the VisualCron service. It will then resume connecting.

I went back to 5.x and it works great without any issue. Went to 6.0.9 again and the issue came back within 10 minutes. It can always be resolved by restarting the VisualCron service in services.msc -- stopping/starting the server via the client did not seem to have any effect.

Any ideas?

It's a HTTP get request with no credentials. No proxy is being used. Exit code is 77777 according to the job log. If I try to debug log it, it just says the request timed out. The HTTP log on apache shows no connection attempt was even made.

Please export the Job and send it to support@visualcron.com so we can test with the exact same settings.

---

Please download TCPView and check if you have a lot of open connections:

http://technet.microsoft...us/sysinternals/bb897437 

---

Yes.. VisualCronService.exe has about 50 connections to the web server in the CLOSE_WAIT state. VisualCron 5.x does not have this issue as far as I can tell as the connections seem to clean themselves up. It appears that v6.x does not clean these up internally?

What happens if you uncheck "Keep alive" on your HTTP Tasks? Then restart service and try to run it again.

Also, are you aborting connections in any way - like using a timeout somewhere?

---

Keepalive is already disabled. There is a 10 second timeout on some of the requests but those requests usually finish in less than 1 second.

I think this is related to timeout. Increase timeout and watch the tcpviewer while it is running to see if connections are cleared automatically.

---

Timeout has been increased to 60 for all http tasks. The problem occurred again about 15 minutes after increasing the timeout. I saw no improvement in tcpviewer, and it seems to be random at how many connections are active as well, regardless of that setting. So right now I have all http tasks at 60 second timeout with keepalive disabled.

1. how many open connections do you have after 15 minutes?
2. normally, if you run the job manually once, is the connection delete right away after the HTTP Task is complete?

---

If run manually, it looks like the connection goes away by itself..
Running for 15 minutes or so runs up about 30-50 connections that are
in time_wait or close_wait state, so not open connections.

Something specific must happens when these connections stay, either at Server or VisualCron level. Timeout, other error etc.

Also, I am guessing that the Apache server has some kind of connection limit. This is the reason for the Request timeout error.

---

I've tested Apache when this occurs. While VisualCron was unable to connect at all, I was able to run apache bench with 100 concurrent connections with 0 failures and also was able to load the URL in my browser. If I go back to 5.8.1 , the issue is was gone for 2 days... I then went to 6.x and it came back within 15 minutes and occurs at least every few hours... any ideas?

Please send log_serverDATE.txt to support@visualcron.com for the date it failed on your server. Please let us know Job and Task name.

---

I have not heard anything back on this and it is still occurring with the latest version. It may even be getting worse.

Please download and install this version:

http://neteject.com/down...on/VisualCron6.1.2-4.exe 

Then go to Server tab->Server settings and uncheck "Override HTTP connection limit". Please give us feedback if this works better.

---