﻿<?xml version="1.0" encoding="utf-8"?><rss version="2.0"><channel><title>Ayende @ Rahien</title><link>http://ayende.com</link><description>Ayende @ Rahien</description><copyright>Copyright (C) Ayende Rahien  2004 - 2021 (c) 2026</copyright><ttl>60</ttl><item><title>Ayende Rahien commented on Things we learned from production, part II&amp;ndash;wake up or I kill you dead</title><description>Rafal,
Given a rich feature set, you need to deal with a lot of variables, and they impact both startup and shutdown.</description><link>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment11</link><guid>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment11</guid><pubDate>Thu, 27 Sep 2012 06:11:26 GMT</pubDate></item><item><title>Rafal commented on Things we learned from production, part II&amp;ndash;wake up or I kill you dead</title><description>Ayende, I was referring to psychological intensity of shutdown and startup procedure, not to the feature set ;) </description><link>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment10</link><guid>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment10</guid><pubDate>Thu, 27 Sep 2012 06:09:31 GMT</pubDate></item><item><title>Ayende Rahien commented on Things we learned from production, part II&amp;ndash;wake up or I kill you dead</title><description>Rafal,
You actually need to do all of that in order to provide a rich feature set.</description><link>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment9</link><guid>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment9</guid><pubDate>Thu, 27 Sep 2012 06:06:30 GMT</pubDate></item><item><title>Ayende Rahien commented on Things we learned from production, part II&amp;ndash;wake up or I kill you dead</title><description>Jesus,
We actually have a slightly more complex behavior now.
We would pick up the phone, ask you to wait, and start things up. If things finish up quickly, we will answer normally.
If things do not finish quickly enough, we will give you an error.</description><link>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment8</link><guid>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment8</guid><pubDate>Thu, 27 Sep 2012 06:05:57 GMT</pubDate></item><item><title>Ayende Rahien commented on Things we learned from production, part II&amp;ndash;wake up or I kill you dead</title><description>Vs,
We make use of some of them, yes. And AppDomain unload is something that we can work with.
What we had a really hard time with was the thread aborts.</description><link>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment7</link><guid>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment7</guid><pubDate>Thu, 27 Sep 2012 06:04:34 GMT</pubDate></item><item><title>oktatás t. commented on Things we learned from production, part II&amp;ndash;wake up or I kill you dead</title><description>this chastity belt is more than expressive:)</description><link>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment6</link><guid>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment6</guid><pubDate>Wed, 26 Sep 2012 20:52:57 GMT</pubDate></item><item><title>Rafal commented on Things we learned from production, part II&amp;ndash;wake up or I kill you dead</title><description>This somehow explains why some posts (like this) are full of pain. I wouldn't trust any software with so rich inner life.</description><link>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment5</link><guid>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment5</guid><pubDate>Wed, 26 Sep 2012 11:37:53 GMT</pubDate></item><item><title>Bob commented on Things we learned from production, part II&amp;ndash;wake up or I kill you dead</title><description>For those who didn't know, the image displays a standard issue underware in the israelian army</description><link>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment4</link><guid>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment4</guid><pubDate>Wed, 26 Sep 2012 10:21:37 GMT</pubDate></item><item><title>Jesús commented on Things we learned from production, part II&amp;ndash;wake up or I kill you dead</title><description>This is my view on what you are/were doing: you receive a phone call at 7:00 am, but you are sleeping. You wake up, but don't pick up the phone, insead you go to have breakfast first, and when you go back to the phone, you have a lost call. You should not do that, instead, you should pick up the phone and suggest her/him call you later because you need to have a cup of coffee first.

IIS calls you, you wake up, but you are not ready to serve the request. IMHO you should not work hard to server the request as soon as possible, instead you should return an error and keep working.

Also I think that after a crash you should recover all databases, but return errors when a client tries to access a not yet recovered database.

SQL Server, for example, behaves this way, it returns you errors just started before databases are recovered, but picks up the phone.</description><link>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment3</link><guid>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment3</guid><pubDate>Wed, 26 Sep 2012 08:46:34 GMT</pubDate></item><item><title>Vs commented on Things we learned from production, part II&amp;ndash;wake up or I kill you dead</title><description>What about CER, CriticalFinalizer, SafeHandle and other things?
And if we talk about AppDomain beign unloaded - I hope you are aware of the fact you can integrate nicely with the WAS runtime?..</description><link>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment2</link><guid>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment2</guid><pubDate>Wed, 26 Sep 2012 06:58:29 GMT</pubDate></item><item><title>tobi commented on Things we learned from production, part II&amp;ndash;wake up or I kill you dead</title><description>Nice idea with moving not-abortable stuff to a separate thread.</description><link>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment1</link><guid>http://ayende.com/158817/things-we-learned-from-production-part-ii-wake-up-or-i-kill-you-dead#comment1</guid><pubDate>Tue, 25 Sep 2012 10:28:55 GMT</pubDate></item></channel></rss>