How can we resolve some of the WinCF server issues? I sometimes only get a “<.” on a coldfusion site. When this happens, non CF sites usually come up, meaning its a “Coldfusion Services Crash”.
Also right now WinCF is “connection was refused…” for the last 10 minutes or more, for all sites.
I have seen the “<.” multiple times a day and wonder how many times those days I missed seeing the error. Even if its one of my sites, can we find out what is going on so it can be fixed?
From: Network Outages
WinCF Reboot
Dated: December 10, 2005
WinCF is undergoing a reboot at the moment. The server will be back up shortly (within 25 minutes)
Saw this after my post, it wasn’t there when I looked before though.
Why a 25 minute reboot?
The server is back up. Regarding your “<” problem, I am quite surprised if you are getting this regularly. We monitor WinCF extensively and a CF crash rarely occurs. Can you PM me your website so we can put it on monitoring
[QUOTE=Yash]
The server is back up. Regarding your “<” problem, I am quite surprised if you are getting this regularly. We monitor WinCF extensively and a CF crash rarely occurs. Can you PM me your website so we can put it on monitoring
[/QUOTE]
This is something that I see too.
We see the <. regularly, at least a dozen times a day. By the time I see it and go to make out a ticket, it has resolved itself.
We see this on www.wheretoliveafter50.com and some of our member-only access sites.
OK, we are going to be monitoring this error. The CF sensor we have isn’t reporting this issue at 60 second intervals. I guess it happens for such a brief period that it cannot pick it up.
I will be putting more sites on monitorin g and shorter intervals so we can catch it
[QUOTE=Yash]
OK, we are going to be monitoring this error. The CF sensor we have isn’t reporting this issue at 60 second intervals. I guess it happens for such a brief period that it cannot pick it up.
I will be putting more sites on monitorin g and shorter intervals so we can catch it
[/QUOTE]
Yash, I don’t know what or which of the monitoring services you use, of the ones that you set up, but I have found that the Windows server services monitor on port 8080 show the problem every time. Then cf server monitors on port 8090 do not show the problem clearly, at least to me.
Whenever I eyeball the problem ocurring while on one of the client sites, I check the monitor on Port 8080 and it shows it everytime. You set this page to refresh every 60 seconds.
BTW, this <. happens when the CF ODBC server restarts itself, CF uses Java and has its own ODBC server, this server randomly resets itself during the day, about once a week it resets and does not start back up, so we are forced to reboot to make that service start.
The reboot takes 25 minutes for a number of reasons, but it really should fully start in under 15, and we are researching further why it takes so long to start all the services.
The <. error is very brief and typically only 1 reporting period long, so the downtime doesn’t show for long unless you visit right as it is happening, as the one reporting period will get buried by many successful periods.
We have some things in the works that will hopefully work out to improve the CFMX6 services, and add more capability in the future.