<br><div class="gmail_quote">On Tue, Jan 27, 2009 at 1:26 PM, Noel Welsh <span dir="ltr"><<a href="mailto:noelwelsh@gmail.com" target="_blank">noelwelsh@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div><br>
</div>Generally our users are in the same time zone, so we can easily<br>
restart services after office hours. Other times we post notice that<br>
the system will be down for a bit.<br>
</blockquote><div><br>I see. I imagine if you have to deal with 5-9s SLA than a backup system would be needed so you can take one mirror down.<br><br></div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
I see these kind of tools as a last resort. Ideally our development<br>
process would catch all bugs before release,. Failing that (and it<br>
will fail) we'd have sufficient logging that we could replay<br>
interactions, and a dev. environment we could debug in, rather than<br>
using the live site. We aren't there yet, but it's the goal.<br>
<font color="#888888"></font></blockquote><div><br>Agreed that such a tool is a last resort. <br><br>I don't want to debug prod env either, and the lack of a dynamic tracing (this is not just a PLT Scheme issue, of course) means that I need to plan ahead to address the situation when my dev process fail. <br>
<br>... need to think about this some more... <br><br>Thanks,<br>yc<br><br></div></div>