Our Website Server Has a Problem
The horrible feeling of clicking on a page, and the site reports a server error, and stays down… Since just after midday yesterday the Apache server hosting graphicline.co.za and all our sub sites has experienced problems. Major configuration changes were rolled out starting on Sunday April 16, which have severely disrupted the function of these sites…
Memory Allocation Reset to Default
First off all, memory allocations were reset to the server default value of 32MB – totally inadequate to run Drupal and WordPress. Then today I discovered sub-domains with static HTML files only were also throwing up a server error – so the sub-domains weren’t being seen as such by the server.
To test a theory, I added a new sub-domain, with a single index.html file and guess what, it wouldn’t serve! What more can I say – the Apache server httpd file is not configured correctly…
During today, I have been watching things, memory gets dropped to 64MB, site fails, gets reset to 100MB or whatever and site comes back… htaccess file gets replaced with one with a single default Apache handler line of script, gets restored to the correct one… Don’t these guys know how a Drupal or WordPress site works? As I write the memory setting is back to 64MB – site down, changed again site up, changed – down, at least 3 times in the last few hours… Come on guys, memory is CHEAP install more if it’s a problem – you cannot run a multi-site server with CMS installations on 16GB of RAM, no matter what the salesman told you – I run more than that on my PC. 96G is what is needed . (I’ll even sell you the memory modules at my cost price).
It’s Not as if Apache Server is Difficult
This is what I don’t get: Apache configuration is not exactly a difficult task – read the software manual! When you know what the problems are, fix them and leave them alone, don’t keep resetting the darn thing to something that cannot possibly work. Why do I have to keep mailing you the same thing over and over. I don’t get paid to tell you how to do the job.
While writing this post (with a food break in between), the sites have been up, down and up again… I don’t even want to log into the Drupal site or WordPress sites admin right now – all I need is a failed database write or something ridiculous to crash the DB – and a Drupal DB is not the quickest thing to fix if the site is completely down with a database error – I know – it was one of the very first things I had to do with a Drupal site – talk being thrown in the deep end.
Watching and Wondering
So for now I’m just watching the server stats, watching the fiddling with script files, and wondering “do these guys really earn a paycheck for holding down a tech support job”. Not from me they wouldn’t. And thinking; “maybe I should mail Rian (the boss) and suggest he hires one person – me – to sort the thing out”. But that is only thought. I am quite happy running my own show, and have no intention of going back to working for someone else – especially any company with a bunch of directors (been there, done that – not my scene at all).
Maybe if this hasn’t been sorted by sometime tomorrow I may suggest they call me in to fix the thing for them – and then they can pay my hourly asking rate too…
First Major Issue in 18 months
My sites have been hosted with this service for 18 months; this is the first time there have been any serious problems so I guess I cannot really complain. The service has recently undertaken confguration work with the server infrastructure. One possible reason the process has had hiccups is these machines are located in a secure data centre managed by a tier one provider, and physical access to the equipment can be difficult. Not everything can be done by remote control.
At least (from their point of view) it was a tech nerds sites (me) it happened to, not some poor customer with no understanding of the difficulties in working on a shared server with numerous live sites.
I Want My Own Server: (or a fishing boat) Yeah, likely