Proving a concept.. HTTP

Portal Home > Knowledgebase > Articles Database > Proving a concept.. HTTP

Posted by appworld, 08-23-2007, 08:55 AM
Would very much appreciate assistance here, I know just enough about Linux to be slightly dangerous and am coming at this issue from a business point of view not technical so excuse anything I say thats technically inept. On Sunday 5th August at 9am, I attempted to access our e-commerce site admin panel to check on sales. Nothing. 404.I tried the main site. Still nothing. I placed a call to Tech support and waited. Let's just ignore the customer service disaster that followed but it wasn't until 9.30pm that evening that I actually spoke to someone. They informed me that Apache had failed due to an oversized log file. I'm ok at this point, I understand what has happened. The technician moved the log file and then brought Apache back up and everything returned to normal. What I'm not sure about is that our 'provider' should provide a fully managed server with service monitoring. 5 point, including HTTP. They are now claiming that HTTP was not down for 12 hours and that their monitoring did not fail but that it was something else that caused me (or 3k customers who usually view my site on a Sunday) to be unable to view our pages. They can't tell me what it is but are adamant that HTTP was running. This is where I need help. a) Am I not correct in thinking that without Apache, there is no serving of web pages? b) That any monitoring system should have recognised this and alerted an engineer? c) How do I go about proving my point? I have root access.. I think.. I have a feeling its not full root. Are there any obvious steps, I would need very clear instructions if its to work, that I can take to prove whether I'm right or wrong? If not, is there anyone who comes highly recommended who can help me? Despite HTTP being down for over 12 hours and no pages being served, the 'provider' is being unhelpful. Any help gratefully recieved! Nicki
Posted by derek.bodner, 08-23-2007, 09:49 AM
If you got a 404 error code, Apache was running. Are you positive it was a 404 and not an "Unable to connect to server"? Often times they will simply ping the server and make sure it is online. They will also simply telnet to port 80 and make sure apache is running. That doesn't do much good if mysql's down or you get a 404, because despite the error, you can still telnet to port 80. If their monitoring software is worth anything, they should be able to search for a keyword on your page, that way you know an error page is not being displayed. Use something that isn't dynamic content, i.e. your copyright is often a place to start.
Posted by ServerNinja, 08-23-2007, 09:55 AM
a) Am I not correct in thinking that without Apache, there is no serving of web pages? Web server (Apache) is responsible for serving the web pages. If you have another web server running on the same server, the web pages can still be accessible. >> b) That any monitoring system should have recognised this and alerted an engineer? A good monitoring system would recognize and alert the engineer if the process goes down. Again, there can be exceptions depending on the method the monitoring system uses (eg: port monitoring, remote plugins etc) >> c) How do I go about proving my point? I have root access.. I think.. I have a feeling its not full root. Are there any obvious steps, I would need very clear instructions if its to work, that I can take to prove whether I'm right or wrong? If not, is there anyone who comes highly recommended who can help me? I would suggest you to hire a server administrator to analyze it.
Posted by plumsauce, 08-24-2007, 02:59 AM
At this point you need representation if you want to pursue the matter. The only way that your hoster is going to concede a point is if they think that they are dealing with someone with technical expertise. Probably not worth the bother. Just pick everything up and move.
Posted by appworld, 08-24-2007, 05:16 AM
Whilst I agree to some point, they are attempting to hold me to 90 days notice, not to mention the thousands of pounds worth of sales lost that day. Profit lost + advertising prices for 13 hours + 3 months of hosting = over £3k. It's not an inconsequential amount of money.
Posted by appworld, 08-24-2007, 05:43 AM
Based on the above.. is there anyone the WHT'ers recommend for a one off investigation? If needs be, we'll go to court on this as I really am that unhappy so I can just include the server investigation prices in the claim.
Posted by david510, 08-24-2007, 06:58 AM
You need to have a check of the apache logs on the server to see the apache was running and was serving pages. You can check the top log as /var/log/dcpumon to see the process that were running during the time apache was not working for you. Not found error can happen when the configuration file got corrupted and wrong conf file was loaded. If this is the case apache will be running but wrong pages will be served when domains are taken in browser.
Posted by jshpik1, 08-24-2007, 09:24 AM
I've seen 404's happen when there was a syntax error in the .htaccess file, so if you have developers I would question them as well.
Posted by david510, 08-24-2007, 11:27 PM
Normally syntax error throws 500 internal server error. If the rule is not correct but a faulty one, it may show 404 error.

Add to Favourites Print this Article