| |
|
Monday, May 6, 2002
|
|
|
I'm working on the new underpinnings for Radio's Instant Outliner. It feels very right to have the pings float over instant messaging. I'm using Eric Soroos's new tool that implements AIM.
4:15:09 PM 
|
|
|
Testing 1, 2, 3...
1:45:46 PM
|
|
|
Business Week: Lawrence Lessig: The "Dinosaurs" Are Taking Over. They've succeeded in making Washington believe this is a binary choice -- between perfect protection or no protection. No one is seriously arguing for no protection. They are arguing for a balance that avoids the phenomenon we are seeing now -- one where the last generation of technology controls the next generation of industry.
11:53:39 AM
|
|
|
NY Times: Access to Free Online Music Is Seen as a Boost to Sales. Disputing the position held by the major record companies, a report issued on Friday found that people who use file-sharing networks to obtain music at no charge over the Internet are more likely to have increased their spending on music than are average online music fans.
11:53:38 AM
|
|
|
This is a test post to help me develop the multi-author weblog stuff.
11:46:40 AM
|
|
|
To Daniel Berlinger: "I think I can shed some light on your queries about centralization and decentralization."
11:46:39 AM
|
|
|
Centralization in blogging tools. First, I'd like to thank Burning Bird for elevating the level of discussion about centralization in weblog tools. As I wrote on Scripting News this morning, we can cover the tools ourselves, as long as we do it scientifically, and with an appreciation for diversity. This is a product category that's just waking up. There's a dearth of real comparative data. It's not about good-vs-bad, it's time to learn and advance the art of weblogging software.
Anyway, I've been weblogging for quite a few years, and I've tried out a lot of the ideas Shelley talks about. I had a discussion group, I've not done comments. I'm watching other people run comment-based sites, and participating where the rules allow it. (Shelley posted a rule a few weeks ago that excludes me from participating in her comments, and of course that's her right, and I am respecting that by commenting here on my Radio weblog.)
I of course think community features are great. I love reading referers lists, for my sites, and others. It helps me ponder the flow of this thing we call the Web. I like knowing what are the most popular RSS feeds. The inter-relation of sites is a never-ending fascination for me. It's the stuff that Jon Udell is writing about. And Josh Allen, the WebMonkey reviewer who organized our little industry so well, is writing about that too.
The community is an important aspect to blogging, imho. We're trying out a new method of Internet-based discourse that's different from the ones that came before. In "blogspace" respect is possible. The weblog is the proxy for a person. We choose the level at which we want to invite others into the space. I like distance, always have -- it helps people be respectful, and that means we can learn new stuff. This discussion of centralization in weblog tools is a perfect example of that. Had we been doing this on a newsgroup or a mail list or discussion group, the Stop Energy would likely overwhelm the intelligent discussion.
So anyway, here's a question for Shelley. When I see your site update on Weblogs.Com, I usually go for a visit to see what the bird is burning about now. I think of that as a community feature. Do you think it's valuable? If not, why do you participate?
11:46:39 AM
|
|
|
Getting bytesInUse back in synch with the content. This is a test. Back in a minute with some notes, if it worked. (Or not.)
Boy that took a lot longer than I thought it should. Here's what's going on.
After the baling-wire-workaround we had a bit of cleanup work to do. With the fix in place we couldn't keep track of bytes in use, because the static server was inaccessible to the community server. So the per-user bytesInUse attribute could be wrong. Now we have to fix them.
Looping over the folder to set bytesInUse used to be fast, a long time ago, before Radio 8 shipped. Now it take a few seconds. First I wrote a utility script, started running it, and realized that it would take a long time to complete. So I opted for a different approach, I did a getUserTable callback, so that when a user connected, either through an upstream or just a ping, the bytesInUse att would get fixed. But the slowness caused timeouts. Quickly backed out of that approach.
I went back to the utility script, and modified it so that it can be restarted. It'll probably take about 10 hours to run. It waits five seconds between users so it can't take up too much of the server's resources while it's running.
I'm going to keep my eye on it today, while I do other cleanup chores and start digging some new holes.
11:46:38 AM
|
|
|
Outage cleared. The lights are back on, with one caveat. We're now running Apache on Windows to do our static serving. Midway through the transition it locked up and we had to restart the machine. Of course since we serve at the pleasure of Murphy, the machine failed to restart. Why? We had installed a new 120GB hard drive, and W2K got lost on its way to restart. Once we cured that, the system started, Apache did its job, and the static server has been fast and reliable since. (It's only been a bit over an hour.) So we will pray for more good luck, it's really gone quite well today, net-net, and we're thankful for the lessons we've learned so far, and look forward to learning many more. 
11:46:36 AM
|
|
|
Test post #2.
11:46:36 AM
|
|
|
Test post.
11:46:35 AM
|
|
|
Outage update, 8AM. 
We have a new static server, it's fully installed at Exodus, with a copy of the content taken overnight. As with the last restoring, if you updated between last night and the time the switch is flipped (probably around 11AM today), some updates will not be there. Rather than having ten hours of downtime, we opted for this approach. We'll provide instructions on updating UserLand-hosted Radio sites after we flip the switch.
The rest of the day today, Murphy-willing, will be spent looking for breakage, and fixing stuff, and undoing the baling-wire and scotch tape we put in place to work around the outage.
I posted a note last night on the Radio-Dev list explaining the strategy for those who are curious and technical.
Dave
11:46:35 AM
|
|
|
WebMonkey: "Radio manages to create a dynamic environment for the exchange of information without asking too much of each individual user. They've made it simple for beginners to get involved in a kind of active network that would've required much more know-how a few years ago. If you're looking for more than just a tool, but an effortless way to get a site launched and incorporated into an online community, Radio may be your best bet."
11:46:34 AM
|
|
|
Outage update, 3:16PM. At 1:10PM we turned off the static server because we were pretty sure it had been hacked, for two reasons:
1. We had received a notice from myNetWatchman with a list of attempts the machine had made to compromise other systems that had the monitoring software installed.
2. We had just finished installing a new sharepoint, and as we were preparing to reconfigure the content system to write through that point, Apache stopped responding, and we saw a flood of traffic coming out of the machine, consistent with what the email said the machine was doing.
At that point we shut it down, and I posted an outage report on Scripting News.
It appears to have been a false alarm for two reasons:
1. Lawrence had the presence of mind to read the email carefully, and noted that all the dates were before we had done the switch to the new server. So the email contained no new information. It was simply telling us that the old server had been hacked. We already knew that. ;->
2. Then we re-examined our assumption that the current static server had been hacked and decided it was worth a test to see, if we backed off the new share, if the machine would go back to its previous performance. It did.
While you can never assume that you're in the clear, and we serve at the pleasure of Murphy, it appears that the static server is working, and we're assume it it was our newbieness with Linux that caused the outage, because it's running OK without the share.
We had a Plan B, which we are now getting ready to execute. It will take a few hours, but we're really optimistic about it.
Dave
11:46:33 AM
|
|
|
IBM: Web services need Sun. In an about-face, Big Blue is paving the way for its rival to join a Web services standards organization, hoping to quell infighting that could put a damper on the emerging technology. [CNET News.com]
11:46:32 AM
|
|
|
Exec exodus continues at Sun. A senior executive responsible for bringing Sun Microsystems closer to open-source technology has left, the latest in a series of high-profile departures for the company. [CNET News.com]
11:46:31 AM
|
|
|
On a wish and a prayer. Ladies and gentlemen, boys and girls..
I was able to string some baling wire, with the help of lots of scotch tape, and now have a very temporary workaround to the outage.
It will be glitchy, for sure, but as you can see -- I am now able to post to my UserLand-hosted Radio weblog.
This buys us a little bit of time now to get the real outage cleared.
Pray to Murphy for our little community.
11:46:31 AM
|
|
|
© Copyright 2002 Jake Savin.
Last update: 5/6/02; 1:25:24 PM.
|
|
| May 2002 |
| Sun |
Mon |
Tue |
Wed |
Thu |
Fri |
Sat |
| |
|
|
1 |
2 |
3 |
4 |
| 5 |
6 |
7 |
8 |
9 |
10 |
11 |
| 12 |
13 |
14 |
15 |
16 |
17 |
18 |
| 19 |
20 |
21 |
22 |
23 |
24 |
25 |
| 26 |
27 |
28 |
29 |
30 |
31 |
|
| May Jun |
|
|
