Hi, I've noted the ZF Wiki's uptime is less than ideal and that in another post Bill Karwin intends to upgrade the whole thing when time permits. It was down again yesterday. I've seen it down for hours at a time.
I'd like to suggest a quick and dirty one liner solution to ease the pain. Something like this in the server's root crontab: 0,15,30,45 * * * * result=`wget --tries 1 --server-response --timeout 60 http://framework.zend.com/wiki/display/ZFDEV/Home 2>&1 | grep 'HTTP/'`; [[ ${result} == " HTTP/1.1 200 OK" ]] || ( echo "Wiki site down. ${result} reported. Attempting restart"; /etc/init.d/apache2 restart ) The above assumes *nix like of course, with wget installed and system v style restart of Apache2. Also, that when a downtime event occurs, the output is emailed to root by a suitable redirect in /etc/aliases to go to a real person. Note the two spaces in the " HTTP/1.1" comparison needed to be there when I tested on another website The restart could be chopped out if it's a worry and at least an email notification would still be sent. I only included the restart as selfishly speaking I'm in the UK and so seem to notice the downtime more when you guys arn't up and about in the US. Hopefully the Wiki upgrade will solve the uptime issue, but I hope people would agree it's no bad thing in principle to have an automated process watching over important web sites' availability. I cannot help think it's a bad thing for newcomers to encounter part of the site down so often as version 1 approaches final release. I understand a robust web site watcher set up would be much more than a crontab one liner, featuring at least two remote watchers travelling over different backbones with back off logic regarding the restart and notification which of course would SMS everyone who works at Zend as well as email them. Not to mention the satellite integration! :) But right now, perhaps the suggested one liner, or a simple script might be more achievable? Any thoughts? Regards, Mark Maynereid
