X040801A: Bad Atom - Diary & Job Jar

Status	Date	Description




	2004-08-27	On the pages where the feed messages are situated (their <id>-elements name those pages), some additional information about how they are customized, the trick about numbering in title and id, and how they disappear when a feed goes back to normal operation needs to be described. When a feed is working, I can actually post an element of that kind, but it would then lead to a permalinked page in the posts section, and that is more than I want, I think.
	2004-08-27	I thought that I could make separate little files that were the "Danger" and "All Clear" announcements, but that would have them disappear when the feed was restored, so that is a no go. I have to continue to do it the hard way!
done	2004-08-26	Post the standby announcement, correct it, and then post the all clear. [dh:2004-08-26 Again, we had new titles and ids for all entries so that they would be recognized as new and in a new time interval. After that, it all worked.]
done	2004-08-26	Make another Danger: Test in Progress announcement and get that fed first. [dh:2004-08-26 We had to make a new title and id (using #'s) so that the feed entry would be recognized as a new incident and also not be an update to an earlier one.
done	2004-08-26	Make new feed standby announcement to be introduced at the half hour.
done	2004-08-26	Improve Actions colorization to work smoothly like just done with the ODMA license project.
failed	2004-08-26	We are going to retest using slamdown pages from a superior directory that are placed using the WS_FTP Copy function. We want to confirm that the datestamp is updated and that the operation appears to be atomic. We will document the advantages of all this later. [dh:2004-08-26 This effort failed. Something didn't work between WS_FTP, my firewall, and the bCentral site, where the copy set up to go and then just didn't happened, with a complaint about no socket from one end or the other. So we are back to a renaming technique.
	2004-08-26	Include a Lessons Learned section at the end of this and related incident reports that provide opportunity for insights. Maybe this should be the key entry on all incident reports!
	2004-08-24	Leave further actions open until site operation is restored and other priorities are satisfied.
done	2004-08-24	Create a feed entry template based on the firedrill/air-raid warning that provides an all-clear from the test.
done	2004-08-24	Confirm that the feed retrieves properly in NewsGator - run the test procedure:
done	2004-08-24	1. Update the site status and all other pertinent information
done	2004-08-24	2. Put the danger message into the wing-atom feed, confirm its retrieval by NewsGator
done	2004-08-24	3. Swap the wing-atom.standby.xml file into place and confirm retrieval by NewsGator
done	2004-08-24	4. Conduct assessment
done	2004-08-24	5. Post the All-Clear message either way
done	2004-08-24	Create a clean wing-atom.standby.xml message that is the same as the official feed except for the special entry.
done	2004-08-24	Create a feed entry template based on the firedrill warning text and insert it in the Wingnut feed.
	2004-08-24	Site pages and the feeds themselves need to provide links to workarounds and status information in case of a bad atom in the barrel. Figure out how to roll that in. People need to have this already, before an incident causes the feed to be lost.
	2004-08-24	Create a special feed for confirming what the required Atom information is, removing all possible dependencies on Blogger for successful access to the feed. This is important because one reason to put an announcement in the feeds is if there is a failure or loss of access to Blogger.
	2004-08-24	Consider putting a feed on the Site Status page for announcing when there are changes to the site that matter. This might be an autodiscovery element, but there should be a button for it too. Put this in an appropriate job jar.
done	2004-08-24	Start the incident report with basic information.