The problem today was report by me in live chat as it happened. So tech support new about it instantly.
We should not have to suffer at the hand of another site that has brought win2 to it’s knee numerous times before.
Not to mention the fact that with many problem I see on various Jodo server, live chat doesn’t even know about it until I tell them. Then it usually takes some time to convince them that I am right. Tighter tolerance must be put in place on the reporting software. And site that have mutiple instances of creating problems for every customer on the server should be removed and suspended immediately.
Perhaps if this wasn’t such a big important 4 letter domain these steps would have been taken already.
You are right. The readouts were ‘spiking’ all over the place - now you would think that:
1, It would set an alarm of some where so the techs could have a look see
2, The server logs would be pulled out for that period automatically so that stephen or some other poor so and so could see where the problem was - with out having to slog thru all the associated data.
3, IF it is the infamous ‘four letter site’ it should have been zapped by now - even if they are paying a lot - Jodo just needs to look at One lost customer vs the other 100s who don’t grow in size cos of bad service, thus kill jodos own growth for the future
We can not always see issues with monitoring software, even at 30 and 60 second intervals in multiple locations inside and outside the network, it is not perfect, also taken into account that sometimes only one app pool(windows 2003) is having an issue, if you have problems simply state the domain and the problem in a livechat, if a problem is happening regularly on one domain ask that we start monitoring that site with the monitoring, so that they get alerts if something happens.
Incidentally, I did raise this via live chat at the time - actually it was probably about 20 mins after the worst of the slow down. However I didn’t get long to discuss it, as live chat seemed to crash (firefox 1.5)
Do I need to set up my own automated site checking system, or should these spikes be picked up by yourselves ? It is a bit worrying, as I for one am not going to wait 11 seconds for anyones web page to load.
ps. not actually annoyed - just want to be sure my site is running as smooth as possible
Cheers
The server is no way overloaded. Infact underloaded. There is no CPU abuse but there seems to be some occasional network abuse that is causing a slowdown in response times. This is extremely hard for us to diagnose, but Stephen has disabled a probable site today and we are monitoring it
Just to proactively update this, there was a very minor slowdown this morning we were right on top of it as it was happening and it lasted all of maybe 3 minutes, page loads however during this time were still about 6 seconds, longer than we want, but not near as bad as it was before.
This is not a very typical issue. We are going through all the logs, everything to find whats the source of this sudden slowness.
right now we have isolated a .NET 2.0 domain that gives us errors in the log just before IIS and everything goes crazy. We have disabled that domain and continue to monitor.
Yes we know, and posted in the status forums, we actually had alarger issue today than in previous days as far as the resolution, as there was an added cause.
Someone was doing asp.net 2.0 DEVELOPMENT(please don’t do this onlive servers, please), every single slowdown today was within 30 seconds of an error thrown by this one .net 2 site. We have alarge number of .net 2 sites, and this is the first time that there have been log entries so close to each slowdown, large or small. This is a newer site on the server, so it was not the cause before 48 hours ago at all, but it added to the issue today. And has been disabled for the time being while we notice them of the problem, and ask to do development off server.
I have found that some asp.net 2.0 code is causing this, one DNN 4.0 site is not loading, and at the same exact time I load it, the server goes haywire.
since I can replicate this problem, I know that it is more than just a mere coincidence.
Well, I am doing one more IIS reset as I found the issue, and made it play up again, I have disabled asp.net 2 on this server for now and am monitoring, some asp.net 2 seems to be working flawlessly, but other times it is causing huge problems. I am researching what the casue of this could/would be, and will re-enable the .net 2 sites as I find out why this is the case.
There are a number of .net 2 sites, but less than the whole server, and the server being stable is more important than the ASP.NET 2 sites at this current time.
[QUOTE=Stephen]
Well, I am doing one more IIS reset as I found the issue, and made it play up again, I have disabled asp.net 2 on this server for now and am monitoring, some asp.net 2 seems to be working flawlessly, but other times it is causing huge problems. I am researching what the casue of this could/would be, and will re-enable the .net 2 sites as I find out why this is the case.
There are a number of .net 2 sites, but less than the whole server, and the server being stable is more important than the ASP.NET 2 sites at this current time.
I’m not sure how many servers have .NET 2, but will you be shutting down .NET 2 on all of them that run it? I will be looking to develop some .NET 2 stuff, but don’t want to get overly excited about it if Jodo will not running it for a while or if there are some real issues. I was wanting to use DNN 4.01, but again not at the cost of instability.
We have disabled .NET 2.0 as a temporary measure only. Response times are fluctuating widly right now, and we are trying to isolate the cause starting from the most obvious things.
.NET 2.0 will be restored and JodoHost is fully committed to providing these services. Win13 is scheduled to go up in another 48 hours which will also be .NET 2.0 enabled.
.NET 2.0 is only disabled on Win2 as we work to diagnose the issue. If .NET 2.0 is responsible, we are committed to finding out what is making it crash
