Cf Update

[Spirited]

Fall out

Now will you show us patience and some assistance getting our sites back on line. I had uploaded some new files, for a site I was going to launch this week, when the original big failure. Not knowing what was going on during all that time, I tried to continue working through that. I know in hindsight I should have went on vacation. But, the biggest fall out for me was that my scheduled events were being lost. That I wasn't aware of, and resulted chasing, fixing problems that didn't exists. Apparently I induced some coding problems. You folks sent me an "abuse" email about my code.

The problem on my end is that I need the robust error reporting to find the problem. This is what got me in trouble during the outage time. The errors are too general. All you get are clueless errors messages. During that time the problem probably wasn't my code, but how do you tell, when you have multiple DB calls - your scheduler isn't working, the servers are being restarted, etc. Now I have a mess and need to get out of the hole.

With errors so general I can only trial and error creating a lot more errors for you log.

I also sent an email into support asking that my schedulers be re-setup and got a message back of some unknown error when you attempted to setup up 2 of the 3 schedulers. Same instructions as these were originally setup under. And, the schedulers were re-setup during the outage when they would get lost. ???


Thanks,

Nick

[Spirited]

Originally Posted by Spirited

Fall out

I also sent an email into support asking that my schedulers be re-setup and got a message back of some unknown error when you attempted to setup up 2 of the 3 schedulers. Same instructions as these were originally setup under. And, the schedulers were re-setup during the outage when they would get lost. ???

A little egg is on my face for this part. I had made one of those long night mistakes of transposing numbers

[Curt]

Spirited, please rememeber that Vortech's servers are not staging servers. If a site is still in development, you should really be working on it on your own system(s). It is common courtesy in a shared environment such as this to limit the chances of your bugs affecting other users. (I'm not by any means suggesting your code had anything to do with the recent CF downtime, by the way. I'm sure it didn't.) I don't know if using Vortech's servers to debug is related to the "abuse" email you received, but it's something to keep in mind for the future. Having your own development systems would've allowed you to continue working right through the outage without any trouble too.

[Spirited]

Curt,

I think you misread. Cuz you sure took off on a tangent to condescend quick.

I wasn't developing on the server. Development was over. Luck would have it that I uploaded the pages the day the CF problems occurred. Problems immediately cropped up. That often happens, the normal thing is too assume your coding is the problem. There was a time when we didn't realize there was a problem on the servers. Then there were the periods when the problem was thought to be fixed. There were the times the servers were being almost constantly restarted without our knowledge.

But, what turned out to be the more devastating problem in my case was the fact schedulers had disappeared and were disappearing most times Vortech performed actions to correct the problem. That fact was not made public and was certainly not as obvious as the DB problems.

Another issue you may not be aware is schedulers are often used in background tasks and problems in that area are not so readily obvious. I.e, you don't see error pages. As matter of fact there are a ton of problems that don't throw up a big red flag telling you something is wrong. You have to create a lot of testing. Testing is very confusing when the servers are constantly being started and restarted.

There was and is no way to test on a development system the fact that you are loosing something like a scheduler on the life system. I for one didn't have the insight to not touch anything and go on vacation, it will have it fixed in 3 days. Sort of like the problem Vortech experienced, life ain't always so cut and dried.

Nick

[Curt]

My sincerest apologies Spirited. I took the start of your message to mean that you were using untested files on the servers (as well as criticizing Vortech's customer committment). If you re-read your first couple sentences, hopefully you can see how I might've made that mistake.

I do hope that people who are developing directly on the production servers see that message though. On the server that most of my sites are on, I see problems occasionally (through IISTracer) with pages that are obviously test code, and it really bothers me when people don't give due respect to being in a shared environment. So when I saw what I thought was someone publicly stating that they were doing exactly that, I decided to say something about it. I mistook your message for something it wasn't and used the moment to vent about something that didn't apply to you at all. So again, my apologies.

Take care and good luck with the new site!

[mresell]

Curt that is a valid point. Brad, I agree with a form for cf submissions. I thing that will help alot. Otherwise, people later will have to search old post and email to check this. Just thing it would be more self-explanatory.

[admin]

We will be adding something to the ODBC section of the user panel to ask your hosting co. for a CF DSN if this is on a ColdFusion server. I have not worked out all the wording for it yet. Craig may do that. He's a little better with writing them I am.

I also spoke with Lawrence Cassese a Manager at MacroMedia today, I gave them some idea of how to fix the neo file issue, I sent him some sample files that were bad for there testing. They should get back to me on tuesday with an answer of if its something they can do. The idea's I told him were rather simple and was kind of shocked they did not or had not do it already.

[mresell]

Originally Posted by admin

They should get back to me on tuesday with an answer of if its something they can do. The idea's I told him were rather simple and was kind of shocked they did not or had not do it already.

Well, doesn't seem like they think there IS a problem. I call that MS disease

[tkraffty]

are we having the problems again today? client site on CF is showing same *nothing* again...

[nabsUK]

I think it was just a blip... less than 2 mins for me.

[admin]

Nothing going on here really unless java flacked out on there. But it looks to be running ok. We restarted it a few hours ago for adding some DSN. We restart it ever 5 DSN we add to be sure it comes back up, so we don't get something poping up 24 hours later.