Picture this:
7am Saturday morning. Kids woken you up. Time to make that big update to your website. Discover that parts of the site aren't working. Database not found, oh but it is for some functions. Perhaps its my browser cache? Flush it very thoroughly. Nope. Oh hell; have breakfast 9am. 10am; see site at local library PC. Duff there too. 11am (back home). Aha; is it where one script (php) calls another? Run phpinfo (shows all the settings), and line-by-line compare with old one. Spot new php version yesterday Friday. Getting warmer? Find setting for global variables changed. Hotter? Look on php.net - yup that's it! And a workaround too (thanks). 4pm - all going again.
This is when I cheerfully want to put a rope around my hosting techies and string them up from the top of the London Eye (big wheel thing next to River Thames).
Now for a positive idea! This is my draft protocol/rules for my hosting company. Is there anything I should add? Thanks.
1. an impact assessment is done for each change proposed (OS, dbms and hardware) and emailed to relevant users
2. you test changes on a test copy of one or more client systems
3. you give email notice of proposed implementations
4. you agree to delay implementations if problems are possible
5. you only make implementations with tech support in place, including over weekends
www.bricksandbrass.co.uk
Reply With Quote
Bookmarks